Targeted compositions for treating or preventing stroke and uses thereof

Abstract

The present application relates to the technical field of biological medicine, and particularly relates to a targeted composition for treating or preventing cerebral stroke and application thereof. The present application provides a brain-targeted composition, which comprises a brain-targeted nucleic acid composition or a brain-permeable small molecule composition; the brain-targeted nucleic acid composition comprises Oct4 gene, Sox2 gene, Klf4 gene, Glis1 gene and Lin28 gene; the brain-permeable small molecule composition comprises the following components at the following concentrations: 150-250 muM valproic acid, 6-10 muM CHIR99021, 6-10 muM Repsox-E616452, 3-5 muM tranylcypromine, and 30-50 muM forskolin. The two compositions are used to intervene in a cerebral stroke model, which can effectively reverse the aging, and provide a basis for the development of drugs for treating cerebral stroke.

Description

Technical Field

[0001] This invention relates to the field of biomedical technology, and in particular to a targeted composition for the treatment or prevention of stroke and its application. Background Technology

[0002] Currently, standard treatments for acute ischemic stroke (such as intravenous thrombolysis and endovascular thrombectomy) are only effective within a very short time window and cannot address the fundamental pathological processes that follow stroke, such as neuroinflammation, cellular senescence, and regenerative disorders, resulting in most patients suffering from severe sequelae.

[0003] In recent years, partial reprogramming technology has offered new insights into reversing cell fate and aging. Studies have shown that widespread cellular senescence-like epigenetic alterations, known as "post-ischemic senescence," exist in the brain after stroke, hindering endogenous repair. Theoretically, reversing the senescent state of these cells through partial reprogramming holds promise for restarting their plasticity and repair functions. However, applying partial reprogramming technology to stroke treatment faces unprecedented challenges: Extreme safety requirements: Brain tissue is highly sensitive to abnormal proliferation and tumorigenic transformation. The carcinogenic risk of the classic reprogramming factor c-Myc is amplified intracranially, making its use absolutely impossible.

[0004] The lack of brain-specific efficacy and safety standards means that the "safety window" observed in the whole body or skin system may not apply to the brain. The brain requires more potent reversals to overcome the epigenetic "barrier" formed by ischemic injury, but is simultaneously more sensitive to over-reprogramming. Currently, the degree to which epigenetic reprogramming can achieve a balance between "maximizing repair" and "minimizing risk" in the brain is completely unknown.

[0005] Lack of assessment methods: The lack of standardized and quantitative methods to accurately measure the epigenetic age changes of specific cell populations in ischemic brain regions makes it impossible to accurately assess the effectiveness of interventions and adjust treatment plans based on objective data.

[0006] Therefore, there is an urgent need in the field for a novel stroke repair strategy designed specifically for brain tissue, which includes a precise quantitative detection method based on authoritative methods (such as the Horvath multi-tissue DNA methylation clock model) and a well-defined safety boundary, and can overcome the aforementioned challenges. Summary of the Invention

[0007] The purpose of this invention is to provide a targeted composition for the treatment or prevention of stroke and its application. A standardized method for quantitative detection of epigenetic age in brain tissue based on DNA methylation clocks will be established to provide core objective indicators for efficacy evaluation.

[0008] To achieve the above-mentioned objectives, the present invention provides the following technical solution: This invention provides a brain-targeting composition, comprising a brain-targeting nucleic acid composition or a brain-penetrating small molecule composition; The brain-targeting nucleic acid composition includes the Oct4 gene, Sox2 gene, Klf4 gene, Glis1 gene, and Lin28 gene; The brain-penetrating small molecule composition comprises the following components at the following concentrations: 150-250 μM valproic acid, 6-10 μM CHIR99021, 6-10 μM Repsox-E616452, 3-5 μM transphenylcyclopropane, and 30-50 μM forscolin.

[0009] Preferably, the genes in the brain-targeting nucleic acid composition are obtained by tandem 2A peptide sequences.

[0010] The present invention also provides the use of the brain-targeting composition in the preparation of medicaments for the treatment and / or prevention of stroke.

[0011] The present invention also provides a brain-targeting expression box, comprising any one of the following: (1) Expression vector: The brain-targeting nucleic acid composition is linked into a delivery vector; (2) Recombinant cells: Brain-targeting nucleic acid composition was transfected into HEK293T cells; The brain-targeting nucleic acid composition is the brain-targeting nucleic acid composition described above.

[0012] Preferably, the delivery carrier is a recombinant adeno-associated virus or modified lipid nanoparticles; The recombinant adeno-associated virus is AAV-PHP.eB or AAV9.

[0013] The present invention also provides the use of the brain-targeting expression cassette in the preparation of medicaments for the treatment and / or prevention of stroke.

[0014] This invention also provides a method for evaluating the effect of non-disease diagnostic drugs on stroke repair, comprising the following steps: (I) Select brain tissue before and after drug intervention and sort out the target cells; (II) Extract genomic DNA from the target cells, perform DNA methylation profile analysis, and obtain methylation data; (III) Input the methylation data into the Horvath epigenetic clock model to obtain the predicted epigenetic age of the target cells; (IV) Substitute the predicted epigenetic age into the formula to calculate the reversal magnitude; (V) Evaluate the repair effect of drugs on stroke based on the degree of reversal.

[0015] Preferably, the drug is the brain-targeting composition described above; The target cells include one or more of neurons, astrocytes, oligodendrocyte precursor cells, and endogenous neural stem cells.

[0016] Preferably, the DNA methylation profiling method is whole-genome bisulfite sequencing or high-precision methylation chip method.

[0017] The preferred formula for calculating the reversal amplitude is as follows: Reversal magnitude % = [(Pre-intervention age acceleration value - Post-intervention age acceleration value) / Pre-intervention age acceleration value] × 100%; Age acceleration value = predicted epigenetic age - actual age; The best effect on stroke repair is achieved when the reversal rate is 57% to 77%.

[0018] The solution of the present invention has the following beneficial effects: (1) Pioneering “detection-treatment” integrated precision system: It deeply integrates a standardized quantitative detection system based on the Horvath epigenetic clock model with the intervention plan, making the treatment process measurable, adjustable and predictable, and realizing true precision medicine.

[0019] (2) Discovery and quantification of brain-specific “safe treatment window”: Through the above detection system, the strict safety boundary (57%~77%) necessary for the regeneration of the central nervous system was objectively defined for the first time, setting an insurmountable quantitative gold standard for therapy development.

[0020] (3) A dynamic balance between repair efficacy and intracranial safety is achieved: By utilizing the feedback of the detection system, the reversal effect of the brain-targeted nucleic acid composition (OSKGL) scheme can be precisely anchored within the treatment window in real time, thereby simultaneously achieving optimal neurological function recovery and zero brain tumor risk.

[0021] (4) Provides a complete solution from diagnosis to treatment: This invention is not only a therapy, but also a complete technical system that includes authoritative efficacy monitoring (based on the Horvath epigenetic clock model) and safety early warning, which greatly improves the feasibility and reliability of clinical translation. Detailed Implementation

[0022] The following detailed description of the solutions provided by the present invention, in conjunction with the embodiments, should not be construed as limiting the scope of protection of the present invention.

[0023] The accession numbers and functions of the Oct4, Sox2, Klf4, Glis1, and Lin28 genes in this embodiment of the invention are shown in Table 1: Table 1 Accession numbers for each gene

[0024] The tandem coding sequence of the five genes is: Oct4-T2A-Sox2-P2A-Klf4-E2A-Glis1-T2A-Lin28. The total polycistronic expression cassette length is approximately 6,800 bp (including the 2A peptide linker). Since the maximum packaging size of a single AAV vector is approximately 4.7 kb, dual AAV vectors are used for packaging.

[0025] The brain-targeting nucleic acid combination of this invention does not contain any nucleic acid sequence encoding the c-Myc protein. c-Myc, as the fourth factor in the classic OSKM protocol, has a highly amplified carcinogenic risk in the intracranial microenvironment, and numerous studies have confirmed its potential to induce gliomas. This invention completely replaces c-Myc with Glis1, fundamentally eliminating this carcinogenic risk while maintaining reprogramming efficiency.

[0026] Auxiliary materials for the induction control system in the embodiments and experimental examples of this invention: (1) Doxycycline hydrochloride (Sigma-Aldrich, D9891): Prepare a stock solution of 100 mg / mL with sterile distilled water and store at -20°C protected from light. Dilute to a final concentration of 2 mg / mL in animal drinking water before use, and add 5% sucrose to improve palatability. Replace the drinking water with fresh Dox every 2 days to prevent degradation.

[0027] (2) Injectors and auxiliary equipment for drug administration: Hamilton microsyringe (100μL, 26G needle) for carotid artery injection of AAV; microsyringe pump (Harvard Apparatus PHD 2000) to control the injection rate of 10μL / min; arterial clamp (miniature, for clamping the pterygopalatine artery and external carotid artery).

[0028] The brain-penetrating small molecule composition described in this invention is a chemical reprogramming cocktail that can achieve partial reprogramming of brain cells without the need for exogenous transcription factor genes. Each component, through synergistic regulation of epigenetic modifications and key signaling pathways, functionally mimics the partial reprogramming effect of the brain-targeting nucleic acid composition (OSKGL): Valproic acid (VPA), as a class I / II histone deacetylases (HDAC) inhibitor, increases histone acetylation levels, opening chromatin accessibility to pluripotency-related loci, functionally equivalent to the early role of Oct4 and Sox2 in initiating chromatin remodeling; CHIR99021, as a highly selective GSK-3β inhibitor, stabilizes β-catenin to activate the Wnt signaling pathway, mimicking Oct4's activation of upstream signal input in the pluripotency transcription network; Repsox (E-61645)... 2) As a selective inhibitor of the TGF-β type I receptor ALK5, it promotes mesenchymal-epithelial transition (MET) by blocking TGF-β / Smad signaling, functionally equivalent to the promoting effect of Klf4 on MET; transphenylcyclopropane (TCP), as an irreversible inhibitor of lysine-specific demethylase 1 (LSD1 / KDM1A), maintains gene activation markers by preventing the demethylation of H3K4me1 / 2, functionally equivalent to the epigenetic activation effect of Glis1; forsocolin, as a direct activator of adenylate cyclase, activates the PKA-CREB signaling cascade by increasing intracellular cAMP levels, enhancing the synergistic reprogramming efficiency of the above four compounds. The combination of the above five compounds achieves partial epigenetic age reversal of brain cells after ischemia through a dual mechanism of epigenetic remodeling (chromatin opening + histone modification reset) and signaling pathway regulation (Wnt activation + TGF-β inhibition + cAMP increase) without introducing any exogenous nucleic acids, and all five compounds have blood-brain barrier permeability, making them suitable for systemic administration. The brain-penetrating small molecule composition and the brain-targeting nucleic acid composition constitute two independent alternatives to the present invention: the nucleic acid composition scheme achieves potent partial reprogramming by directly expressing the OSKGL transcription factor, which is suitable for severe stroke scenarios requiring deep epigenetic reversal; the small molecule composition scheme achieves mild epigenetic regulation by chemically mimicking reprogramming signals, without gene delivery-related risks, and is suitable for clinical scenarios requiring higher safety or where gene therapy is unavailable.

[0029] Example 1

[0030] Composition and preparation of brain-targeting nucleic acid composition (OSKGL)

[0031] 1.1 Composition of the nucleic acid composition

[0032] After codon optimization, the genes in Table 1 were tandemly linked into polycistronic expression cassettes using self-cleaving 2A peptide sequences (T2A / P2A / E2A).

[0033] The amino acid sequence of T2A is shown in SEQ ID NO:1, and the nucleotide sequence is shown in SEQ ID NO:2; the amino acid sequence of P2A is shown in SEQ ID NO:3, and the nucleotide sequence is shown in SEQ ID NO:4; the amino acid sequence of E2A is shown in SEQ ID NO:5, and the nucleotide sequence is shown in SEQ ID NO:6. SEQ ID NO: 1: EGRGSLLTCGDVEENPGP; SEQ ID NO: 2: GAGGGCAGAGGAAGTCTGCTAACATGCGGTGACGTTGAGGAGAATCCTGGACCT; SEQ ID NO:3: GSGATNFSLLKQAGDVEENPGP; SEQ ID NO: 4: GGAACGGAGCTACTAACTTCAGCCTGCTGAAGCAGGCTGGAGACGTGGAGGAGAACCCTGGACCT; SEQ ID NO:5: QCTNYALLKLAGDVESNPGP; SEQ ID NO: 6: CAGTGTACTAATTATGCTCTCTTGAAATTGGCTGGAGATGTTGAGAGCAATCCCGGGCCT.

[0034] 1.2 Composition of Control Element

[0035] To achieve time-controlled expression of the brain-targeting nucleic acid composition (OSKGL) in the brain, a tetracycline-inducible (Tet-On) system was employed, specifically composed of the following regulatory elements: (1) Promoter system: TRE3G (third generation tetracycline response element), containing 7×tetO sequence repeats and a minimal CMV promoter, driving the OSKGL polycistronic expression cassette. This promoter remains silent in the absence of doxycycline (Dox) and achieves efficient transcription after Dox activates the trans-activator.

[0036] (2) Transactivator: rtTA3M (reverse tetracycline-controlled transactivator, triple mutant optimized), is composed of a reverse mutant of E. coli TetR protein fused with the VP16 transcriptional activation domain. Compared with wild-type rtTA, rtTA3M has lower basal leakage activity (approximately 10-fold reduction) and higher Dox sensitivity (EC). 50(Approximately 100 ng / mL). rtTA3M is continuously expressed by either a neuron-preferred human promoter or a broad-spectrum CAG promoter.

[0037] (3) Safety braking element: TetR-KRAB transcriptional repressor. When Dox is withdrawn, TetR-KRAB binds to the tetO sequence on TRE3G, actively silencing the promoter region and recruiting heterochromatin remodeling, forming a "double insurance" shutdown mechanism (Dox withdrawal → rtTA3M inactivation + TetR-KRAB active silencing), ensuring that the reprogramming factor is completely shut down after the treatment window ends.

[0038] 1.3 Design and Preparation of AAV Carriers

[0039] The chosen serotype for the vector is AAV-PHP.eB. This serotype was selected by the Bejamine Deverman laboratory at Caltech using the CREATE technique. Its capsid protein contains a 7-peptide insert sequence (TLAVPFK), enabling it to efficiently cross the blood-brain barrier (BBB). After tail vein injection, its transduction efficiency to neurons and astrocytes in the brain is approximately 40-55 times higher than that of AAV9. For local delivery scenarios that do not rely on systemic administration, the AAV9 serotype can also be used.

[0040] Dual AAV split-intein packaging strategy: Since the total length of the OSKGL polycistronic protein exceeds the single AAV packaging limit, the expression cassette is split into two AAV vectors, and the Npu DnaE split-intein system is used to achieve protein trans-splicing recombination in cells. AAV-OSKGL-N (carrier 1, approx. 4.5 kb): 5'ITR–TRE3G–Oct4-T2A-Sox2-P2A-Klf4(N-terminus)-IntN–WPRE–BGH PolyA–ITR3'.

[0041] AAV-OSKGL-C (carrier 2, approx. 4.4 kb): 5'ITR–TRE3G–IntC-Klf4 (C-terminal)-E2A-Glis1-T2A-Lin28A–WPRE–BGH PolyA–ITR3'.

[0042] AAV-rtTA3M (carrier 3, approx. 3.2 kb): 5'ITR––rtTA3M–WPRE–BGH PolyA–TetR-KRAB–SV40 PolyA–ITR3'.

[0043] IntN and IntC are the N-terminal (123aa) and C-terminal (36aa) fragments of the Npu DnaE inteptide, respectively. When the two vectors are co-transduced into the same cell, IntN and IntC spontaneously bind, precisely reconstructing the full-length Klf4 protein through protein splicing, while releasing the complete OSKGL multiprotein precursor, which generates five independent functional proteins after cleavage of the 2A peptide.

[0044] Specific preparation method of AAV vector: Step 1, Plasmid Construction: The three expression cassettes were cloned into the pAAV-MCS backbone plasmid (Addgene, catalog number #46954, containing the AAV2 inverted terminal repeat sequence and the ampicillin resistance gene). Cloning was performed using the Gibson Assembly method, with 15-25 bp homologous overlapping sequences designed at the ends of each fragment. Assembly was completed in one step using Gibson Assembly Master Mix (New England Biolabs, catalog number E2611) under isothermal conditions at 50°C via T5 exonuclease, Phusion DNA polymerase, and Taq DNA ligase. Specifically, the assembly fragments of vector 1 (AAV-OSKGL-N) include: a pAAV-MCS backbone linearized with NheI / HindIII, a TRE3G promoter fragment, an Oct4-T2A-Sox2-P2A-Klf4 (N-terminus)-IntN coding fragment, and a WPRE-PolyA fragment; the assembly strategies for vectors 2 (AAV-OSKGL-C) and 3 (AAV-rtTA3M) are the same. The assembly products were transformed into Stbl3 competent cells (Thermo Fisher Scientific) to avoid recombination instability of the ITR sequence in conventional DH5α. Single clones were selected, identified by restriction endonuclease digestion, and positive clones were sent for Sanger sequencing to verify the correctness of the full-length sequence, including ITR integrity, reading frames of each gene, 2A peptide linker sequences, split-intein splicing sites, and the orientation of regulatory elements.

[0045] AAV2 is shown in SEQ ID NO: 7; SEQ ID NO: 7: CCTGCAGGCAGCTGCGCGCTCGCTCGCTCACTGAGGCCGCCCGGGCAAAGCCCGGGCGTCGGGCGACCTTTGGTCGCCCGGCCTCAGTGAGCGAGCGAGCGCGCAGAGAGGGAGTGGCCAACTCCATCACTAGGGGTTCCT.

[0046] The TRE3G promoter is as shown in SEQ ID NO:8; SEQ ID NO:8: GAGTTTACTCCCTATCAGTGATAGAGAACGTATGTCGAGTTTACTCCCTATCAGTGATAGAGAACGATGTCGAGTTTACTCCCTATCAGTGATAGAGAACGTATGTCGAGTTTACTCCCTATCAGTGATAGAGAACGTATGTCGAGTTTACTCCCTATCAGTGATAGAGAACGTATGTCGAGTTTACTCCCTATCAGTGATAGAGAACGTATGTCGAGTTTATCCCTATCAGTGATAGAGAACGTATGTCGAGTTTACTCCCTATCAGTGATAGAGAACGGTATGTCGAGTTTACCCTATCAGTGATAGAGAACGTATGTCGAGCTCGGTACCCGGGTCGAGTAGGCGTGTACGGTGGGAGGCCTATATAAGCAGAGCTCGTTTAGTGAACCGTCAGATCGC。

[0047]

[0048] The amino acid sequence of Oct4 is shown in SEQ ID NO:10, SEQ ID NO:10: MAGHLASDFAFSPPPGGGDGSAGLEPGWVDPRTWLSFQGPPGGPGIGPGSEVLGISPCPPAYEFCGGMAYCGPQVGLGLVPQVGVETLQPEGQAGARVESNSEGTSSEPCADRPNAVKLEKVEPTPEESQDMKALQKELEQFAKLLKQKRITLGYTQADVGLTLGVLFGKVFSQTTICRFEALQLSLKNMCKLRPLLEKWVEEADNNENLQEICKSETLVQARKRKRTSIENRVRWSLETMFLKCPKPSLQQITHIANQLGLEKDVVRVWFCNRRQKGKRSSIEYSQREEYEATGTPFPGGAVSFPLPPGPHFGTPGYGSPHFTTLYSVPFPEGEAFPSVPVTALGSPMHSN.

[0049] The CDS of Sox2 is shown as SEQ ID NO:11, SEQ ID NO:11: ATGTATAACATGATGGAGACGGAGCTGAAGCCGCCGGGCCCGCAGCAAGCTTCGGGGGGCGGCGGCGGAGGAGGCAACGCCACGGCGGCGGCGACCGGCGGCAACCAGAAGAACAGCCCGGACCGCGTCAAGAGGCCCATGAACGCCTTCATGGTATGGTCCCGGGGGCAGCGGCGTAAGATGGCCCAGGAGAACCCCAAGATGCACAACTCGGAGATCAGCAAGCGCCTGGGCGCGGAGTGGAAACTTTTGTCCGAGACCGAGAAGCGGCCGTTCATCGACGAGGCCAAGCGGCTGCGCGCTCTGCACATGAAGGAGCACCCGGATTATAAATACCGGCCGCGGCGGAAAACCAAGACGCTCATGAAGAAGGATAAGTACACGCTTCCCGGAGGCTTGCTGGCCCCCGGCGGGAACAGCATGGCGAGCGGGGTTGGGGTGGGCGCCGGCCTGGGTGCGGGCGTGAACCAGCGCATGGACAGCTACGCGCACATGAACGGCTGGAGCAACGGCAGCTACAGCATGATGCAGGAGCAGCTGGGCTACCCGCAGCACCCGGGCCTCAACGCTCACGGCGCGGCACAGATGCAACCGATGCACCGCTACGACGTCAGCGCCCTGCAGTACAACTCCATGACCAGCTCGCAGACCTACATGAACGGCTCGCCCACCTACAGCATGTCCTACTCGCAGCAGGGCACCCCCGGTATGGCGCTGGGCTCCATGGGCTCTGTGGTCAAGTCCGAGGCCAGCTCCAGCCCCCCCGTGGTTACCTCTTCCTCCCACTCCAGGGCGCCCTGCCAGGCCGGGGACCTCCGGGACATGATCAGCATGTACCTCCCCGGCGCCGAGGTGCCGGAGCCCGCTGCGCCCAGTAGACTGCACATGGCCCAGCACTACCAGAGCGGCCCGGTGCCCGGCACGGCCATTAACGGCACACTGCCCCTGTCGCACATG。

[0050] The amino acid sequence of Sox2 is shown in SEQ ID NO:12. SEQ ID NO:12: MYNMMETELKPPGPQQASGGGGGGGNATAAATGGNQKNSPDRVKRPMNAFMVWSRGQRRKMAQENPKMHNSEISKRLGAEWKLLSETEKRPFIDEAKRLRALHMKEHPDYKYRPRRKTKTLMKKDKYTLPGGLLAPGGNSMASGVGVGAGLGAGVNQRMDSYAHMNGWSNGSYSMMQEQLGYPQHPGLNAHGAAQMQPMHRYDVSALQYNSMTSSQTYMNGSPTYSMSYSQQGTPGMALGSMGSVVKSEASSSPPVVTSSSHSRAPCQAGDLRDMISMYLPGAEVPEPAAPSRLHMAQHYQSGPVPGTAINGTLPLSHM.

[0051]

[0052] The amino acid sequence of Klf4 is shown in SEQ ID NO:14, SEQ ID NO:14: MRQPPGESDMAVSDALLPSFSTFASGPAGREKTLRPAGAPTNRWREELSHMKRLPPLPGRPYDLAATVATDLESGGAGAACSSNNPALLARRETEEFNDLLDLDFILSNSLTHQESVAATVTTSASASSSSSPASSGPASAPSTCSFSYPIRAGGDPGVAASNTGGGLLYSRESAPPPTAPFNLADINDVSPSGGFVAELLRPELDPVYIPPQQPQPPGGGLMGKFVLKASLTTPGSEYSSPSVISVSKGSPDGSHPVVVAPYSGGPPRMCPKIKQEAVPSCTVSRSLEAHLSAGPQLSNGHRPNTHDFPLGRQLPTRTTPTLSPEELLNSRDCHPGLPLPPGFHPHPGPNYPPFLPDQMQSQVPSLHYQELMPPGSCLPEEPKPKRGRRSWPRKRTATHTCDYAGCGKTYTKSSHLKAHLRTHTGEKPYHCDWDGCGWKFARSDELTRHYRKHTGHRPFQCQKCDRAFSRSDHLALHMKRHF。

[0053] The nucleotides of Npu DnaE IntN are shown in SEQ ID NO:15, SEQ ID NO:15: TGCCTGTCCTACGAGACAGAGATCCTGACCGTGGAGTACGGCCTGCTGCCAATCGGCAAGATCGTGGAGAAGCGGATCGAGTGCACCGTGTACTCAGTGGACAACAACGGCAACATCTACACCCAGCCCGTGGCCCAGTGGCACGACAGAGGCGAGCAGGAGGTGTTCGAGTACTGCCTGGAAGATGGCTCCCTGATCCGGGCCACCAAGGACCACAAGTTCATGACAGTGGACGGCCAGATGCTGCCCATCGACGAGATCTTCGAGAGAGAGCTGGACCTGATGAGAGTGGACAACCTGCCCAACTCAGGCGACCAGCTGACCCAGGAAGATGAGATCAAGAACTTCGTGAAGCAGAAGATCGATAGACTGAACGAGGAG。

[0054] The nucleotides of Npu DnaE IntC are shown in SEQ ID NO:16, SEQ ID NO:16: TGCTTCAACAAGGAGATCACCATCCACAACGTGTCCGAGTCCACCAAGCAGCTGGACATCGCCAACCAGCTGCTGAAGATCGAGTACGAGCGGAATGTGAAGTTTGAGATCGACTCCAGA。

[0055] The nucleotides of WPRE are shown in SEQ ID NO:17, SEQ ID NO:17: AATCAACCTCTGGATTACAAAATTTGTGAAAGATTGACTGGTATTCTTAACTATGTTGCTCCTTTTACGCTATGTGGATACGCTGCTTTAATGCCTTTGTATCATGCTATTGCTTCCCGTATGGCTTTCATTTTCTCCTCCTTGTATAAATCCTGGTTGCTGTCTCTTTATGAGGAGTTGTGGCCCGTTGTCAGGCAACGTGGCGTGGTGTGCACTGTGTTTGCTGACGCAACCCCCACTGGTTGGGGCATTGCCAC。

[0056] The nucleotides of BGH polyA are shown in SEQ ID NO:18, SEQ ID NO:18: CTGTGCCTTCTAGTTGCCAGCCATCTGTTGTTTGCCCCTCCCCCGTGCCTTCCTTGACCCTGGAAGGTGCCACTCCCACTGTCCTTTCCTAATAAAATGAGGAAATTGCATCGCATTGTCTGAGTAGGTGTCATTCTATTCTGGGGGGTGGGGTGGGGCAGGACAGCAAGGGGGAGGATTGGGAAGACAATAGCAGGCATGCTGGGGATGCGGTGGGCTCTATGG。

[0057] The CDS (C-terminal) of Klf4 is shown in SEQ ID NO:19, SEQ ID NO:19: TGCCTGCCAGAGGAGCCCAAGCCAAAGAGGGGAAGAAGGTCGTGGCCCCGGAAAAGAACAGCCACCCACACTTGTGACTATGCAGGCTGTGGCAAAACCTATACCAAGAGTTCTCATCTCAAGGCACACCTGCGAACTCACACAGGCGAGAAACCTTACCACTGTGACTGGGACGGCTGTGGGTGGAAATTCGCCCGCTCCGATGAACTGACCAGGCACTACCGCAAACACACAGGGCACCGGCCCTTTCAGTGCCAGAAGTGTGACAGGGCCTTTTCCAGGTCGGACCACCTTGCCTTACACATGAAGAGGCACTTT。

[0058]

[0059] The amino acid sequence of Glis1 is shown in SEQ ID NO:21. SEQ ID NO:21: MHCEVAEALSDKRPKEAPGAPGQGRGPVSLGAHMAFRIAVSGGGCGDGNPLDLLPRLPVPPPRAHDLLRPRSPRDYGVSKTGSGKVNGSYGHSSEKSLLDLDLAEGPSPSCHQGLFLPAGTPPPRGHPPVCEKLLHFPHPNRSPRPQATFVNGSLPAAQHIKQEALPDYQAMVSAHTPLPTHCRAPSSMGLPSDLDFPDRGLTNPAPSCYLLGNEPISDLGPQPEAHLPEGSLKRCCLLGLPPTSSASSSPCASSDINPVIHSSQTALVSCVNGLRSPPLPGDLGGPPKRSRPGPASSDGQEGSLQLEACRKSGFLKQEPMDEFSELFAPHHQGLPPPYPLPQLPTGPGLGGLGLGLAGRMVAGRQACRWVDCCAAYEQQEELVRHIEKSHIDQRKGEDFTCFWAGCVRRYKPFNARYKLLIHMRVHSGEKPNKCMFEGCSKAFSRLENLKIHLRSHTGEKPYLCQHPGCQKAFSNSSDRAKHQRTHLDTKPYACQIPGCSKRYTDPSSLRKHVKAHSAKEQQVRKKLHTGADPEADVLSECLSLQQLQASTLLPASRGKGSQTLSQELLPGVYPGSVTPQNGLASGILSPSHDVPSRHHPLEVPTGSHHHLSPLPTAESTRDGLGPSLLSPMVSPLKGLGPPPLPPASQSQSPGGQSFSTVPSKPTYPSFQSPPPLPSPQGYQGSFHSIQNCFPYADCYRATEPAASRDGLVGDAHGFNPLRPSTYSSLSTPLSAPGYETLAETPCPPALQPQPAEDLVPSGPEDCGFFPNGAFDHCLSHIPSIYTDT。

[0060] The nucleotides of the Lin28A CDS are shown as SEQ ID NO:22, SEQ ID NO:22: ATGGGCTCGGTGTCCAACCAGCAGTTTGCAGGTGGCTGCGCCAAGGCAGCGGAGAAGGCGCCAGAGGAGGCGCCGCCTGACGCGGCCCGAGCGGCAGACGAGCCGCAGCTGCTGCACGGGGCCGGCATCTGTAAGTGGTTCAACGTGCGCATGGGGTTCGGCTTCCTGTCTATGACCGCCCGCGCTGGGGTCGCGCTCGACCCCCCGGTGGACGTCTTTGTGCACCAGAGCAAGCTGCACATGGAAGGGTTCCGAAGCCTCAAGGAGGGTGAGGCGGTGGAGTTCACCTTTAAGAAGTCTGCCAAGGGTCTGGAATCCATCCGTGTCACTGGCCCTGGTGGTGTGTTCTGTATTGGGAGTGAGCGGCGGCCAAAAGGGAAGAACATGCAGAAGCGAAGATCCAAAGGAGACAGGTGCTACAACTGCGGTGGGCTAGACCATCATGCCAAGGAATGCAAGCTGCCACCCCAGCCCAAGAAGTGCCACTTTTGCCAAAGCATCAACCATATGGTGGCCTCGTGTCCACTGAAGGCCCAGCAGGGCCCCAGTTCTCAGGGAAAGCCTGCCTACTTCCGGGAGGAAGAGGAAGAGATCCACAGCCCTGCCCTGCTCCCAGAAGCCCAGAAT。

[0061] The amino acid sequence of Lin28A is shown as SEQ ID NO:23, SEQ ID NO:23: MGSVSNQQFAGGCAKAAEKAPEEAPPDAARAADEPQLLHGAGICKWFNVRMGFGFLSMTARAGVALDPPVDVFVHQSKLHMEGFRSLKEGEAVEFTFKKSAKGLESIRVTGPGGVFCIGSERRPKGKNMQKRRSKGDRCYNCGGLDHHAKECKLPPQPKKCHFCQSINHMVASCPLKAQQGPSSQGKPAYFREEEEEIHSPALLPEAQN。

[0062] The nucleotides of the CDS of rtTA3M are shown in SEQ ID NO:24, SEQ ID NO:24: ATGTCCCGGCTGGACAAAAGCAAAATCATTAACTCTGCCCTCGAGCTGCTCAATGGCGTCGGCATTGAGGGACTGACTACCAGAAAGTTGGCCCAAAAGCTGGGCGTCGAGCAACCCACTCTCTATTGGCATGCCAAAAACAAACGGGCTTTGCTGGACGCTCTGCCTATCGAAATGTTGGACAGACACCATACCCATAGCTGTCCCCTGGAGGGAGAGTCCTGGCAAGACTTTCTGAGAAACAATGCCAAAAGCTATCGGTGTGCCTTGCTGTCCCACAGAAACGGAGCCAAAGTGCATCTCGGAACCAGACCCACTGAGAAACAGTATGAGACTCTCGAAAACCAATTGGCTTTCTTGTGCCAACAGGGATTCTCTCTGGAAAACGCTCTCTATGCCCTCTCTGCTGTGGTCCACTTTACCCTGGGCTGTGTGCTCGAGGAACAGGAACACCAAGTGGCTAAGGAAGAGAGAGAGACTCCCACTACCGATTCTATGCCCCCTCTGCTCAAGCAAGCCATTGAGCTCTTCGATCGGCAAGGCGCTGAGCCTGCCTTTCTGTTTGGCCTGGAGTTGATCATTTGCGGACTCGAAAAGCAATTGAAATGCGAAAGCGGATCTAGCCGGGGAGACGCTTTGGATGACTTTGACCTCGACATGCTCGGAAGCGATGCCCTGGACGATTTCGATTTGGATATGTTGGGCTCCGACGCTCTCGATGACTTTGACCTGGACATGCTGTGA。

[0063] The amino acid sequence of rtTA3M is shown in SEQ ID NO:25, SEQ ID NO:25:MSRLDKSKIINSALELLNGVGIEGLTTRKLAQKLGVEQPTLYWHAKNKRALLDALPIEMLDRHHTHSCPLEGESWQDFLRNNAKSYRCALLSHRNGAKVHLGTRPTEKQYETLENQLAFLC QQGFSLENALYALSAVVHFTLGCVLEEQEHQVAKEERETPTTDSMPPLLKQAIELFDRQGAEPAFLFGLELIICGLEKQLKCESGSSRGDALDDFDLDMLGSDALDDFDLDMLGSDALDDFDLDML.

[0064] The nucleotide of SV40 polyA is shown in SEQ ID NO: 26, SEQ ID NO: 26: AACTTGTTTTATTGCAGCTTATAATGGTTACAAATAAAGCAATAGCATCACAAATTTCACAAATAAAGCATTTTTTTCACTGCATTCTAGTTGTGGTTTGTCCAAACTCATCAATGTATCTTATCATGTCTGGATCTGATCACTGCTTGAGCCTAGGAGATCCAGACATGATAAGATACATTGATGAGTTTGG.

[0065] Step 2, AAV production via three-plasmid co-transfection: AAV was produced in HEK293T cells (ATCC CRL-3216) using an industrial-standard three-plasmid co-transfection method. The three plasmids were: ① pAAV target gene plasmid (vector 1, 2, or 3); ② pHelper helper plasmid (providing E2A, E4, and VA RNA); ③ pAAV-PHP.eB Rep / Cap plasmid (containing the AAV2 Rep gene and the PHP.eBCap gene, Addgene#103005).

[0066] Transfection conditions: HEK293T cells were cultured in DMEM medium containing 10% FBS. When the cell confluence reached 75%, co-transfection was performed using polyethyleneimine (PEI, linear 25 kDa, molecular weight ratio PEI:DNA = 3:1). The molar ratio of the three plasmids was: target plasmid:pHelper:pAAV-PHP.eB Rep / Cap = 1:1:1. 45 μg of total DNA was used per 15 cm culture dish. The medium was changed 6 h after transfection, and the cells were cultured for another 72 h before harvesting.

[0067] Step 3, Crude AAV Extraction: At harvest, adherent cells are scraped and the supernatant is collected. Cells are subjected to repeated freeze-thaw cycles (-80℃ / 37℃, three cycles) to release intracellular AAV particles. Benzoenase (final concentration 50 U / mL, treated at 37℃ for 30 min) is added to digest the free nucleic acids. The cells are centrifuged at 4,000×g for 30 min to remove cell debris, and the supernatant containing AAV is collected to obtain the crude AAV extract.

[0068] Step 4, Iodixanol density gradient ultracentrifugation purification: Prepare a discontinuous iodixanol (OptiPrep) gradient: four layers of 15%, 25%, 40%, and 60%. Load the crude AAV extract onto the top of the gradient and ultracentrifuge at 350,000 × g (Beckman 70Ti rotor) at 18°C ​​for 90 min. Collect the AAV-containing fraction (approximately 4 mL) at the 40% / 60% interface.

[0069] Step 5, buffer replacement and concentration: Transfer the collected AAV fraction into a 100 kDa MWCO ultrafiltration tube (Amicon Ultra-15), use PBS + 0.001% Pluronic F-68 as the replacement buffer, centrifuge at 3,000×g and repeatedly dilute-concentrate 4 times to a final volume of approximately 350 μL.

[0070] Step Six, Titer Determination and Quality Control: ① Genomic Titer: Quantitative PCR (qPCR) is used, employing primer and probe sets targeting WPRE elements or ITR sequences, with quantification based on a linearized plasmid standard curve. Target titer ≥ 1 × 10⁻⁶. 13 ① Purity test: SDS-PAGE silver staining should show three main bands: VP1 (87 kDa), VP2 (73 kDa), and VP3 (62 kDa), with no obvious contaminating proteins. ② Endotoxin test: LAL method, <5 EU / mL. ③ Infectivity titer: Immunofluorescence detection of reporter gene expression (if applicable) or RT-qPCR detection of transgenic mRNA should be performed 72 h after HEK293T cell transduction. ④ Sterility test: Inoculate with TSB and FTM media and incubate at 37°C for 14 days for sterility.

[0071] Final formulation composition: purified AAV-PHP.eB suspension, carrier is PBS (pH 7.4) + 0.001% Pluronic F-68, stored at -80℃. Thaw and briefly centrifuge before use, avoid repeated freeze-thaw cycles more than 3 times.

[0072] 1.4 Dosing regimen

[0073] Carotid artery injection: Three AAV carriers were mixed in equimolar ratios, with a total dose of 3 × 10⁻⁶. 11vg / rat (1×10 of each vector) 11 The veno-carotid artery (vg) was slowly injected via the internal carotid artery in a total volume of 100 μL (injection rate 10 μL / min). The pterygopalatine and external carotid arteries were clamped before injection to ensure that the carrier preferentially entered the cerebral vascular bed.

[0074] Dox induction protocol: Starting on day 3 after AAV administration, doxycycline (2 mg / mL, with 5% sucrose for dissolution) was added to drinking water and administered continuously for 14 days. Dox was then withdrawn, and the TetR-KRAB active silencing program was initiated. This 14-day period is the "therapeutic window" for OSKGL expression.

[0075] Example 2

[0076] Composition and preparation of brain-penetrating small molecule compositions

[0077] 2.1 The detailed composition of the brain-penetrating small molecule composition is shown in Table 2.

[0078] Table 2 Composition of brain-penetrating small molecule compositions

[0079] 2.2 Specific preparation method of brain-penetrating small molecule composition

[0080] Step 1, Preparation of mother liquor for each component: (1) VPA stock solution: Weigh sodium valproate (Sigma-Aldrich, catalog number P4543) and dissolve it in sterile physiological saline to a stock solution of 100 mM. VPA is a water-soluble compound and dissolves at room temperature. After sterilization by filtration through a 0.22 μm filter membrane, dispense and store at 4°C. Shelf life is 6 months.

[0081] (2) CHIR99021 stock solution: Weigh CHIR99021 (MedChemExpress, catalog number HY-10182) and dissolve it in DMSO to a 10 mM stock solution. Vortex until completely dissolved (a brief ultrasonic water bath can be used if necessary). Dispense into light-proof EP tubes and store at -20°C for 12 months. Avoid repeated freeze-thaw cycles more than 5 times.

[0082] (3) Repsox stock solution: Weigh Repsox (MedChemExpress, catalog number HY-13012) and dissolve it in DMSO to a 10mM stock solution, following the same procedure as CHIR99021. Store at -20℃ away from light.

[0083] (4) Transphenylcyclopropylamine (TCP) stock solution: Weigh transphenylcyclopropylamine hydrochloride (Sigma-Aldrich, catalog number P8511) and dissolve it in sterile water to a 10 mM stock solution. Sterilize by 0.22 μm filtration, aliquot and store at -20℃.

[0084] (5) Forscoline stock solution: Weigh out forscoline (MedChemExpress, catalog number HY-15371) and dissolve it in DMSO to a 50 mM stock solution. Store at -20℃ away from light.

[0085] Step 2, Preparation of working solution (intravenous injection formulation): In a sterile operating room, take appropriate amounts of each stock solution and add them to the following basic solvent system to prepare a single-use injection working solution: The base solvent composition is primarily physiological saline (0.9% NaCl), supplemented with 5% 2-hydroxypropyl-β-cyclodextrin (HP-β-CD, Sigma-Aldrich, H107) as a solubilizer (to improve the water solubility of CHIR99021, Repsox, and foscorlin), and 0.5% PEG-400 as a solubilizing stabilizer. The final solution DMSO volume fraction is controlled to ≤1% to ensure safety for intravenous injection.

[0086] The final concentrations of each component were calculated based on in vivo pharmacokinetic conversion (assuming a rat plasma volume of 10 mL / 300 g body weight, and considering the first-pass effect and BBB permeability), and the specific injection doses are shown in Table 3.

[0087] Table 3 Injection Dosage

[0088] Step 3, Quality Control of the Formulation: ① pH Testing: Adjust to pH 7.2 (fine-tune with phosphate buffer). ② Osmolarity Testing: Within the range of 280~320 mOsm / kg. ③ Visible Foreign Matter and Clarity Inspection: Visually, it should be a colorless to pale yellow clear liquid with no visible particles. ④ Endotoxin Testing: LAL method, requirement <5 EU / mL. ⑤ Component Content Testing: HPLC method, the recovery rate of each component should be 90%~110% of the nominal value. ⑥ Sterility Testing: After sterilization filtration at 0.22μm, samples are taken and inoculated into TSB and FTM media for verification.

[0089] Step 4, Stability: Store the prepared working solution at 4°C protected from light. Allow it to return to room temperature before use. It is recommended to use within 24 hours. For long-term storage, freeze at -20°C. Thaw slowly and mix gently before use.

[0090] Example 3

[0091] Construction of an epigenetic age quantitative detection model

[0092] 3.1 Reagents for Brain Tissue Dissociation and Single-Cell Suspension Preparation

[0093] (1) Composition of enzyme digestion solution: Hibernate-A medium as the base solution, with the addition of papain (20 U / mL, Worthington, LS003126), DNase I (100 U / mL, Roche, 10104159001), and L-cysteine ​​(1 mM). It was pre-activated at 37℃ for 15 min before use.

[0094] (2) Termination solution: Hibernate-A medium containing 10% FBS.

[0095] (3) Red blood cell lysis buffer: ACK lysis buffer (150 mM NH4Cl, 10 mM KHCO3, 0.1 mM EDTA, pH 7.4).

[0096] (4) FACS buffer: PBS + 2% FBS + 1 mM EDTA + 0.05% NaN3.

[0097] Brain tissue dissociation procedure (cell suspension): Approximately 40 mg of cortical tissue from a 2 mm periphery of the ischemic penumbra was collected and rapidly minced to approximately 1 mm³ in pre-chilled Hibernate-A solution. The tissue was transferred to enzyme digestion buffer and digested at 37°C with gentle pipetting every 10 min to ensure homogeneity. An equal volume of stop solution was added to terminate the digestion. The cells were filtered through a 70 μm cell sieve and centrifuged at 300 × g for 5 min to collect the cell pellet. The pellet was treated with ACK lysis buffer for 2 min to remove red blood cells, washed twice with PBS, and resuspended in FACS buffer to obtain a single-cell suspension.

[0098] 3.2 Fluorescence-activated cell sorting (FACS)

[0099] The sorted cells are shown in Table 4.

[0100] Table 4 Target Cells

[0101] Since NeuN and Sox2 are nuclear antigens, cells must be fixed and transmembrane-transferred using the BD Cytofix / Cytoperm Fixation and Transmembrane Kit (BD, 554714) before FACS. GFAP is also an intracellular protein and similarly requires fixation and transmembrane transfer. The recommended sorting strategy is as follows: first, gate live cells with live / dead dyes, then sort the target population sequentially using a single marker or a combination of markers, collecting ≥50,000 cells from each population for downstream DNA extraction.

[0102] 3.3 DNA Extraction and Bisulfite Conversion Reagents

[0103] (1) Genomic DNA Extraction: QIAamp DNA Micro Kit (Qiagen, 56304) is suitable for efficient DNA recovery from small numbers of cells after FACS sorting (recovery rate >80%, starting amount as low as 100 cells). Follow the kit's standard operating procedure. Key steps: Proteinase K digestion at 56°C for 1 h (instead of the standard 10 min) to ensure that the cross-linked proteins after fixation and membrane permeation are fully digested. Add vector RNA (1 μg, provided in the kit) to improve the column recovery efficiency of trace DNA. Quantify using the QubitdsDNA HS Assay Kit (Thermo, Q32851), requiring a total DNA volume ≥100 ng (approximately 2 ng / μL × 50 μL). OD260 / 280 = 1.8~2.0. Samples that do not meet the requirements should be re-sorted.

[0104] (2) Bisulfite Conversion: EZ DNA Methylation-Gold Kit (Zymo Research, D5005). This kit combines CT conversion reagent (containing sodium bisulfite and a protectant) with column purification, achieving a conversion efficiency >99% and DNA recovery rate >75%. Key points of operation: Strictly control the denaturation temperature (98℃ for 10 min) and conversion temperature (64℃ for 2.5 h) to ensure complete conversion of unmethylated cytosine to uracil.

[0105] 3.4 DNA Methylation Profiling Methods

[0106] Method 1 (preferred in this embodiment): Simplified Representative Bisulfite Sequencing (RRBS). The standard RRBS workflow is employed, consisting of MspI digestion (recognizing CCGG sites and enriching CpG-dense regions) + end repair + adapter ligation + bisulfite transformation + PCR amplification + high-throughput sequencing (Illumina NovaSeq 6000 or equivalent platform, PE150 mode, ≥30M clean reads per sample). RRBS covers approximately 1-3% of CpG sites (1-3 million) across the entire genome, focusing on CpG islands and promoter regions. It significantly reduces costs compared to WGBS and is suitable for large-scale screening across multiple groups.

[0107] Method 2 (High-Precision Alternative): Whole Genome Bisulfite Sequencing (WGBS). Direct library construction and sequencing after bisulfite transformation, covering >28 million CpG sites across the entire genome. Sequencing depth ≥30×. Suitable for studies requiring comprehensive methylation mapping.

[0108] Method 3 (Chip Replacement): Illumina Infinium Mouse Methylation BeadChip (approximately 285,000 CpG sites), suitable for standardized high-throughput screening. Perform according to Illumina standard operating procedures.

[0109] 3.5 Bioinformatics Analysis Pipeline and Horvath Clock Calculation

[0110] (1) Raw data preprocessing (RRBS / WGBS): Trim Galore was used for adapter removal and quality trimming (Q≥20). The data was aligned to the mouse reference genome (mm10 / GRCm38) using Bismark (v0.24.0 or later) to obtain the methylation β value (number of methylated reads / total number of reads) for each CpG site.

[0111] (2) Horvath Epigenetic Clock Calculation: The β-value matrix (β=M / (M+U+α)) was input into a publicly available mouse-specific multi-tissue Horvath clock model. Specifically, the "DNAmAge" calculation package in R (such as the online calculator released by Steve Horvath's lab) was used. This model is based on 350 validated CpG marker sites and outputs the "predicted epigenetic age" (in weeks) for each sample through a resilient network regression algorithm.

[0112] (3) Calculation of the accelerated age value: Accelerated age value = predicted epigenetic age - actual age in weeks. A positive value represents epigenetic "accelerated aging" (such as after ischemic injury), while a negative value or a value close to zero represents a normal or younger epigenetic age. For example, if a rat with an actual age of 12 weeks has a predicted epigenetic age of 32 weeks for its astrocytes in the ischemic penumbra, then AA = +20 weeks, indicating that ischemic injury has accelerated the epigenetic aging of cells in this region by 20 weeks.

[0113] (4) Calculation of reversal magnitude: Reversal magnitude (%) = [(age acceleration value before intervention - age acceleration value after intervention) / age acceleration value before intervention] × 100%; The meaning of this formula is: if the epigenetic age of astrocytes in a certain MCAO rat before intervention is accelerated by 20 weeks due to ischemia (acceleration value +20 weeks), and the acceleration value drops to +6.4 weeks after intervention, then the reversal magnitude = (20-6.4) / 20×100% = 68%.

[0114] Experimental Example 1

[0115] Validating the "safe therapeutic window for neuroepidermal remodeling" and the efficacy of the OSKGL regimen in a rat MCAO model based on DNA methylation clock.

[0116] I. Experimental Procedure

[0117] 1. Establishment of experimental animals and MCAO model

[0118] 1.1 Laboratory Animals

[0119] Two hundred and ten adult male SPF-grade Sprague-Dawley rats were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. The animals were housed in an SPF-grade barrier environment at a temperature of 22±2°C and humidity of 50±10%, with a 12-hour / 12-hour light-dark cycle and free access to food and water. All animal experiments were approved by the institution's animal ethics committee and strictly adhered to the 3R principle (Responsiveness, Resilience, and Harmony). The rats were introduced into the experiments after a 7-day acclimatization period.

[0120] 1.2 Establishment of a permanent middle cerebral artery occlusion (pMCAO) model

[0121] Surgical procedure selection: A permanent middle cerebral artery occlusion model was established using the modified Longa suture occlusion method. The specific procedure is as follows: (1) Anesthesia and preoperative preparation: Anesthesia was administered via inhalation of isoflurane (5% for induction, 1.5-2.0% for maintenance, with 30% O2 / 70% N2O as the carrier gas). Rats were fixed in a supine position on a 37°C constant-temperature heating pad, and the core body temperature was continuously monitored by a rectal temperature probe to maintain it at 36.5-37.5°C. The surgical area was prepared and disinfected. A PE-50 catheter was inserted through the tail artery to continuously monitor the mean arterial pressure (MAP). Rats with a MAP below 60 mmHg were excluded.

[0122] (2) Exposure of neck vessels: Make a midline neck incision (2 cm), and separate the subcutaneous tissue and muscle layer by layer to expose the left common carotid artery (CCA), external carotid artery (ECA) and internal carotid artery (ICA). Ligate the proximal end of the CCA and the distal branches of the ECA (occipital artery, superior thyroid artery), and apply a temporary arterial clamp proximal to the ECA stump.

[0123] (3) Insertion of the suture: Make a small incision at the ECA stump, and insert a 4-0 nylon suture (diameter 0.36±0.02 mm, length 40 mm, purchased from Doccol Corporation, 403756PK10) with its tip coated with poly-L-lysine into the CCA bifurcation point through the ECA stump. After adjusting the direction, slowly advance it along the ICA until a slight resistance is felt (about 18~20 mm from the CCA bifurcation point). At this point, the tip of the suture is just blocking the origin of the middle cerebral artery (MCA). Fix the suture and leave it permanently without removal (permanent occlusion model).

[0124] (4) Postoperative management: The incision was sutured, and buprenorphine (0.05 mg / kg) was administered subcutaneously for analgesia. The rats were placed in a recovery cage at 32°C until they were fully awake. Six hours after the operation, 1 mL of 5% glucose saline was administered subcutaneously to prevent dehydration.

[0125] Sham surgery group (G1): (5) Sham surgery group operation: completely consistent with the above steps, but the suture plug is only pushed into the ICA for about 10mm and then immediately withdrawn, so as not to cause MCA occlusion.

[0126] 1.3 Model Success Criteria and Exclusion Criteria

[0127] (1) Immediately after surgery (1-2 hours), Longa neurological deficit assessment was performed: 0 points = no neurological deficit; 1 point = unable to fully extend the contralateral forepaw; 2 points = turning to the contralateral side; 3 points = falling to the contralateral side; 4 points = no spontaneous activity and loss of consciousness. The inclusion criteria were Longa scores of 1-3. Those with scores of 0 (model unsuccessful) and 4 (excessive injury) were excluded.

[0128] (2) Exclusion of subarachnoid hemorrhage: Exclusion is given to patients who die within 24 hours after surgery and whose autopsy reveals subarachnoid hemorrhage.

[0129] (3) Expected exclusion rate: approximately 15-20%. 35 patients were initially operated on in each group to ensure that ≥25 patients could be included in the final analysis in each group.

[0130] 2. Grouping and Intervention Program

[0131] Rats with successful pMCAO model (Longa score 1-3) were randomly divided into 6 groups (G2-G7) (Table 5) using a random number table method, plus the sham-operated group G1, for a total of 7 groups. Grouping was completed 24 h postoperatively to ensure that there were no statistically significant differences in baseline neurological function scores among the groups.

[0132] Table 5. Grouping of rats

[0133] 2.1 Detailed administration procedures for each group

[0134] G5 (OSKGL nucleic acid group) (brain-targeting nucleic acid composition group of Example 1): 24 h post-surgery, rats were re-anesthetized while awake to expose the left internal carotid artery. Three AAV vectors (AAV-OSKGL-N, AAV-OSKGL-C, AAV-rtTA3M) were mixed in equimolar ratios, with a total dose of 3 × 10⁻⁶. 11 vg / rat (1×10 of each vector) 11 The dose (vg) was adjusted to a total volume of 100 μL with PBS + 0.001% Pluronic F-68 and slowly injected via the internal carotid artery at a rate of 10 μL / min. The pterygopalatine and external carotid arteries were clamped before injection to ensure that the carrier preferentially enters the cerebral vascular bed.

[0135] After injection, hold the arterial clamp for 5 minutes before releasing and suturing the incision. Starting on day 3 after AAV administration, add Dox (2 mg / mL + 5% sucrose) to drinking water for 14 days (G5) until day 17 when Dox is removed.

[0136] Dox mechanism of action: Dox is activated at D3 → rtTA3M is activated → binds to TRE3G → drives OSKGL polycistronic expression. Dox is removed at D17 → rtTA3M is inactivated (primary shutdown) + TetR-KRAB occupies the tetO site and recruits KRAB-mediated heterochromatin silencing (secondary shutdown), ensuring complete termination of OSKGL expression with double protection.

[0137] Other control procedures: ① Postoperative days 1-14: blank solvent administered via tail vein twice daily (controlling the G6 variable). ② BrdU: same as above.

[0138] G6 (OSKGL small molecule group) (brain-penetrating small molecule group of Example 2): Administration began 24 hours post-surgery, twice daily (bid, 12-hour interval), via tail vein injection for 14 consecutive days. The working solution composition for each injection was: VPA 150 mg / kg + CHIR99021 10 mg / kg + Repsox 8 mg / kg + TCP 5 mg / kg + Forskolin 20 mg / kg, dissolved in physiological saline containing 5% 2-hydroxypropyl-β-cyclodextrin HP-β-CD + 0.5% PEG-400, with a total injection volume of approximately 1.5 mL / 300 g body weight. The injection rate was 0.5 mL / min.

[0139] Safety monitoring: Monitor weight and general condition (activity level, coat color, food and water intake) daily. Rotate injection sites (alternating between left and right tail veins) to reduce vascular damage. Collect tail blood every 3 days to test liver and kidney function (ALT, AST, BUN, Cr) to monitor the systemic toxicity of compounds such as VPA. If ALT is greater than 3 times the upper limit of normal or body weight decreases by more than 15%, discontinue administration for 48 hours and reassess.

[0140] Other control procedures: ① Postoperative day 1: 100 μL PBS injected into the carotid artery (controlling the G5 operating variable). ② Days 3-17: Sucrose drinking water (without Dox, controlling the G5 water intake variable). ③ BrdU: Same as above.

[0141] Rationale for group design: G6 achieves non-viral mediated partial reprogramming through a small molecule chemical composition, representing an alternative therapeutic pathway that does not require gene delivery. The expected reversal rate is 62% ± 9%, falling within the safety window but slightly lower than G5 (because small molecules are not as precise as AAV in terms of sustained expression). The comparison between G6 and G5 validates that both different delivery modalities (gene vs. chemical) can regulate the reversal rate within the safety window, demonstrating the universality of the therapeutic window concept. G6 has significant implications for future clinical translation—small molecule drugs are significantly superior to gene therapy in terms of production, storage, administration, and regulatory pathways.

[0142] G7 (High-intensity OSKGL group): The procedure was the same as G5, with postoperative injection via the internal carotid artery on day 1, but the total AAV dose was increased to 9×10¹¹vg / animal (3×10¹¹vg for each carrier), and the Dox drinking time was extended to 21 days (7 days longer than the standard protocol) to deliberately increase the reversal rate beyond the upper limit of the safety window as a safety control.

[0143] Dox extended induction protocol: D3~D24: Dox was administered via drinking water for 21 days (7 days longer than G5), which prolonged the expression time of OSKGL factor by 50%. TetR-KRAB was initiated to silence the OSKGL factor after Dox was withdrawn on D24.

[0144] The TetR-KRAB consists of a TetR (618bp) connector, a GGSGGGS connector (24bp), and a KRAB (222bp). The nucleotides of TetR are shown in SEQ ID NO:27, SEQ ID NO:27: ATGTCTAGATTAGATAAAAGCAAAGTCATAAATTCTGCTCTGGAACTACTCAACGGAATCGGTATCGAAGGCCTGACGACAAGGAAACTCGCTCAAAAGCTGGGAGTTGAGCAGCCTACCCTGTACTGGCACGTGAAGAACAAGCGGGCCCTGCTCGATGCCCTGGCAATCGAGATGCTGGACAGGCATCATACCCACTCCTGCCCCCTGGAAGGCGAGTCATGGCAAGACTTTCTGCGGAACAACGCCAAGTCATTCCGCTGTGCTCTCCTCTCACATCGCGACGGGGCTAAAGTGCATCTCGGCACCCGCCCAACAGAGAAACAGTACGAAACCCTGGAAAATCAGCTCGCGTTCCTGTGTCAGCAAGGCTTCTCCCTGGAGAACGCACTGTACGCTCTGTCCGCCGTGGGCCACTTTACACTGGGCTGCGTATTGGAGGAACAGGAGCATCAAGTAGCAAAAGAGGAAAGAGAGACACCTACCACCGATTCTATGCCCCCACTTCTGAAACAAGCAATCGAGCTGTTCGACCAGCAGGGAGCCGAACCTGCCTTCTTTTTCGGCCTGGAACTAATCATATGTGGCCTGGAGAAACAGCTAAAGTGCGAAAGCGGC。

[0145] GGSGGGS is shown in SEQ ID NO:28, SEQ ID NO:28: GGCGGAGGATCTGGAGGCGGATCT。

[0146] The nucleotide of KRAB is shown in SEQ ID NO:29, SEQ ID NO:29: ATGGCCGCCTTCACCAAGTGTCCCATCGAGATCAAGGCTCAGGTGCAGAATAGCAAGAACCGCATGATCACCGGCCTGCAGGAGTGCAACAACCAGGATACCAACTACCTGTTCTACCAGATGTACACAAACTGCGAGATCAAGCTGAATGGCGTGGAGACCGAGGGCATGATCCACTTCAAGAACGGCGAGAAGGCCCTGAAGGACCTGATCCCTCTGTAG.

[0147] Other control procedures: ① Postoperative days 1-14: blank solvent administered via tail vein twice daily. ② BrdU: same as above.

[0148] Rationale for group design: G7 is the most critical safety control group in the safety window theory of this invention. Its design aims to intentionally push the epigenetic age reversal rate above the upper limit of the safety window (expected 83% ± 11%, >77%), in order to directly verify the two core hypotheses: a) The harmful hypothesis of excessive reversal: It is expected that 25% of the G7 animals (5 / 20 surviving to the sampling point) will have abnormal proliferative foci or glioma-like lesions in the brain at sampling point D28 or D90, with IHC showing Ki-67 >10%, diffuse strong GFAP positivity, and possibly IDH1-R132H positivity. This would directly prove that reversal exceeding the 77% upper limit is unsafe in brain tissue.

[0149] b) Hypothesis of the early warning function of DNA methylation clock: The Horvath clock of G7 can detect over-reversal of >77% as early as D28, much earlier than tumor formation that can only be clearly diagnosed on histopathology at D90. This proves the clinical value of the detection system of this invention as an early warning tool—the clock has already issued a high-risk signal before the tumor actually appears.

[0150] c) Verification of the accuracy of the therapeutic window boundary: The comparison between G7 and G5 (within the window), together with the data of G4 (below the window), the three groups jointly outline the complete dose-effect-safety curve of the reversal magnitude from insufficient to safe and effective to excessively dangerous, providing solid experimental support for the 57%~77% therapeutic window.

[0151] G3 (positive control group): Starting 24 hours post-surgery, edaravone injection was diluted to a working concentration of 2 mg / mL with 0.9% saline and administered via tail vein twice daily for 14 consecutive days (postoperative days 1 to 14). Edaravone is a stroke treatment drug approved by the China NMPA (trade name: Bicun), and was used as a positive control for standard clinical treatment.

[0152] Other procedures: ① Postoperative day 1: 100 μL PBS injected into the carotid artery (to control surgical procedures). ② Days 3-17: Sucrose-infused water (without Dox). ③ BrdU: Same as above.

[0153] Rationale for group design: Edaravone is an approved stroke treatment (free radical scavenger) widely used in clinical practice. As a positive control group, G3 is significant for: a) validating that the experimental system of this protocol can detect the efficacy of known effective drugs (system sensitivity validation); b) comparing the efficacy of the OSKGL regimen head-to-head with current standard clinical treatment. G3 is expected to show a moderate improvement in mNSS (2-4 points) compared to G2, but its epigenetic age reversal is limited (expected <57%, i.e., below the lower limit of the safety window) because edaravone, as a free radical scavenger, does not possess epigenetic reprogramming capabilities.

[0154] G1 (Sham Operation Group): Dosing Regimen (Complete Solvent Control): To control for placebo effect and procedural consistency, G1 patients received a solvent / carrier control dose that was completely matched to the treatment group. ① Carotid artery injection (simulating AAV administration): Postoperative D1: 100 μL of PBS + 0.001% Pluronic F-68 was injected via the internal carotid artery (10 μL / min bolus injection, same procedure as G5) to simulate the AAV administration process.

[0155] ② Tail vein injection (simulating small molecule drug delivery): Postoperative D1~D14: 1.5 mL / 300g body weight of normal saline containing 5% HP-β-CD + 0.5% PEG-400, administered via tail vein injection, bid, simulating G6 small molecule drug delivery.

[0156] ③ Drinking water (simulating Dox induction): D3~D17: Add 5% sucrose (without Dox) to the drinking water to simulate Dox induction. Replace the drinking water bottle every 2 days.

[0157] ④ BrdU labeling: Postoperative D2~D15: BrdU 50 mg / kg, ip, bid (consistent with all groups), used for subsequent neurogenesis detection.

[0158] Reason for group design: To provide normal baseline values ​​for all detected indicators in the experiment; its GFAP +The epigenetic age acceleration value (AA) of astrocytes is theoretically close to 0 (reflecting only the weak effect of surgical stress) and serves as a baseline for calculating the magnitude of reversal.

[0159] G2 (MCAO model group): Surgical procedure: Performed according to standard pMCAO protocol. A poly-L-lysine-coated 4-0 nylon suture (0.36±0.02 mm in diameter, Doccol 403756PK10) was inserted into the ICA through the ECA stump and advanced to 18-20 mm from the CCA bifurcation point, where resistance was felt, and then fixed. The suture was permanently retained and not removed. Postoperative Longa score of 1-3 was considered.

[0160] Dosing regimen (complete solvent control): Completely identical to G1: ① Postoperative D1: 100μL PBS injected into the carotid artery (same as G1).

[0161] ②Postoperative D1~D14: blank solvent administered via tail vein twice daily (same as G1).

[0162] ③D3~D17: Sucrose drinking water (same as G1).

[0163] ④BrdU: Same as G1.

[0164] Rationale for group design: To provide data on the natural outcome of disease in patients who did not receive any treatment after pMCAO; their epigenetic age acceleration value represents the maximum degree of aging caused by ischemic injury (before intervention) and serves as the denominator for calculating the reversal magnitude of each treatment group. It is expected that their mNSS will remain at a high level (9-14 points) on day 28, with no significant reduction in infarct volume and a significantly elevated SASP factor.

[0165] 3. Quantitative detection of epigenetic age based on Horvath DNA methylation clock

[0166] Complete standardized operating procedures: 3.1 Timing of brain tissue collection Main detection time point: Day 28 after intervention (i.e., Day 29 after pMCAO surgery). Ten rats were randomly selected from each group for epigenetic age detection (they were euthanized after sampling).

[0167] Auxiliary time point: Five additional animals / groups were selected on day 14 after intervention for mid-term epigenetic age monitoring to assess the trend of clock dynamic changes.

[0168] 3.2 Brain tissue acquisition and ischemic penumbra localization

[0169] (1) Rats were deeply anesthetized with an excessive amount of sodium pentobarbital (150 mg / kg, ip) and their brains were removed by decapitation. The brain tissue was immediately placed in ice-cold Hibernate-A medium.

[0170] (2) The whole brain was sectioned coronally using a rat brain mold (Roboz, SA-). Cortical tissue from the region 2 mm lateral to the edge of the infarct core (i.e., the ischemic penumbra) was taken, which is a key area for the dynamic balance of cell senescence and repair. Approximately 40 mg of tissue was taken from each rat for downstream processing.

[0171] (3) In the sham surgery group (G1), the cortical tissue at the corresponding anatomical location was taken as the baseline control.

[0172] 3.3 Preparation of Single-Cell Suspension

[0173] The obtained brain tissue was placed in a pre-cooled 35 mm culture dish, and a single-cell suspension was prepared according to the method in Example 3.

[0174] 3.4 Fluorescence-activated cell sorting (FACS)

[0175] Using GFAP in Example 3 + Astrocytes were the primary detection target (a key senescent cell type after stroke), and NeuN cells were also sorted. + Neurons are used as auxiliary verification.

[0176] (1) Fixation and membrane penetration: Since both GFAP and NeuN are intracellular / nuclear antigens, fixation and membrane penetration are required before sorting. a. First, stain dead cells with Zombie Aqua live / dead dye (BioLegend#423102, 1:500 dilution) at 4°C in the dark for 15 min to mark them. Wash once with PBS.

[0177] b. Add BD Cytofix / Cytoperm fixation membrane-penetrating buffer (BD, 554714) and incubate at 4°C for 20 min for fixation. Wash twice with 1×BD Perm / Wash buffer.

[0178] (2) Antibody staining: Add the following antibody combination to BD Perm / Wash buffer and incubate at 4°C in the dark for 30 min. The antibodies are shown in Table 6.

[0179] Table 6. Antibody Sources

[0180] (3) Sorting strategy and operation: Use BD FACSAria III or an equivalent high-speed sorter. The gating strategy is as follows: FSC / SSC gated debris removal → single-cell gating (FSC-H vs FSC-W) → Zombie Aqua negative (live cells, since cells that died before fixation are excluded in this step) → GFAP + Door (AF488 channel) or NeuN + Gating (AF647 channel). Collect ≥50,000 cells from each population into 1.5 mL centrifuge tubes containing 300 μL PBS. Include fluorescence minus one (FMO) and isotype control tubes to determine the gating threshold. Target purity of ≥95% for sorting.

[0181] 3.5 Genomic DNA Extraction

[0182] (1) Centrifuge the sorted cells (5,000×g, 5 min, 4°C), discard the supernatant, and extract genomic DNA according to the method in Example 3.

[0183] 3.6 Bisulfite sequencing (RRBS)

[0184] RRBS, as a DNA methylation profiling method, offers high coverage of CpG islands and promoter regions, is moderately costly, and is suitable for experimental designs involving multiple groups and samples. The specific workflow is as follows: Step 1: MspI digestion. 100 ng of genomic DNA was digested with the MspI restriction endonuclease (NEB, R0106). MspI recognizes CCGG sites, and the resulting fragments are enriched with CpG-dense regions. Reaction system: DNA 100 ng + MspI 20 U + CutSmart Buffer 1×, digested overnight (16 h) at 37°C. Enzyme inactivation was performed at 65°C for 20 min.

[0185] Step 2: End repair and A-tail addition. Use Klenow Fragment (3′→5′exo) - (NEB, M0212) End-completion and A addition were performed under conditions containing dATP, dCTP, and dGTP (excluding dTTP). Incubation was performed at 30°C for 20 min, followed by at 37°C for 20 min. Purification was achieved using AMPure XP magnetic beads (1.8× ratio).

[0186] Step 3: Methylation adapter ligation. Using methylation adapters (Illumina TruSeq or NEBNext methylation adapters), ligation was performed overnight at 16°C under the catalysis of T4 DNA Ligase (NEB, M0202). Adapter dimers were removed by purification with AMPure XP magnetic beads (1× ratio).

[0187] Step 4: Bisulfite Conversion: The conversion was performed according to the bisulfite conversion procedure in Example 3. Under these conditions, the conversion rate of unmethylated C→U was >99.5%, and the retention rate of methylated C was >99%. After column purification, the solution was eluted with 10 μL of M-Elution Buffer.

[0188] Step 5: PCR Amplification and Library Purification. PCR amplification was performed using PfuTurbo Cx Hotstart DNA Polymerase (Agilent, 600414; this enzyme reads uracil and does not produce PCR bias). Program: 95°C 2 min, (95°C 30 s, 60°C 30 s, 72°C 45 s) 14 cycles, 72°C 7 min. The number of cycles was adjusted according to the amount of starting DNA to avoid over-amplification (gel electrophoresis should show diffuse 150-500 bp smear-like bands). Double-sided purification was performed using AMPure XP magnetic beads (0.8× scale) to select 150-500 bp library fragments.

[0189] Step 6: Library quality control. The distribution of library fragments (peak value approximately 250–350 bp) was detected using an Agilent 2100 Bioanalyzer or Agilent TapeStation. The library concentration was accurately quantified using the KAPA Library Quantification Kit (Roche, KK4824) qPCR method.

[0190] Step 7: High-throughput sequencing. After mixing equal volumes of libraries from each sample, sequence them on an Illumina NovaSeq 6000 platform in PE150 mode. Each sample should produce ≥30M clean reads (approximately 4.5 Gb) to ensure sufficient CpG site coverage (average >10×).

[0191] 3.7 Bioinformatics Analysis and Horvath Clock Calculation

[0192] After the raw data were assessed for quality by FastQC, the predicted epigenetic age, age acceleration value, and reversal magnitude were calculated using the bioinformatics analysis pipeline of Example 3 and the Horvath clock.

[0193] The quality trimming parameters were -quality 20 -length 36 -rrbs -paired. The bowtie2 parameter was used when aligning to the mouse reference genome. The methylation level β value of CpG was calculated as methylated reads / total reads, filtering out CpG sites with a coverage depth <5×.

[0194] 3.8 Calculation of Age Acceleration Value and Reversal Amount

[0195] The aging acceleration value and reversal magnitude were calculated according to the formula in Example 3.

[0196] 4. Neurological function evaluation

[0197] 4.1 Modified Neurological Deficit Score (mNSS)

[0198] mNSS was the primary efficacy endpoint. Scoring was performed simultaneously by two independent assessors with completely blinded grouping information, and the average score was taken. Assessment time points: postoperative day 1 (baseline), day 7, day 14, and day 28.

[0199] The mNSS has a total score of 18 points, which includes the following four sub-items, as shown in Table 7.

[0200] Table 7 Detection Sub-items

[0201] Total score interpretation: 0 points = normal; 1~6 points = mild injury; 7~12 points = moderate injury; 13~18 points = severe injury. Inclusion criteria: mNSS baseline score of 7~14 points on postoperative day 1 (moderate to moderate to severe injury). Patients with scores that are too low (<7) or too high (>14) were excluded to reduce heterogeneity.

[0202] 4.2 Rotard Test

[0203] Assisted motor function assessment indicators. A Ugo Basile 47750 accelerator rotator was used. For the first 3 days pre-surgery, training was conducted 3 times daily (acceleration from 5 to 40 rpm, completed within 5 minutes), and the fall latency of the last training session was recorded as the baseline. Post-surgery, tests were conducted 3 times each on days 7, 14, and 28, and the longest fall latency (in seconds) among the three tests was recorded as the performance at that time point. Rotator speed scheme: initial speed 5 rpm, uniformly accelerated to 40 rpm within 300 seconds. Maximum recording time 300 seconds.

[0204] 4.3 Morris Water Maze (MWM, only available for 28 days)

[0205] Assess spatial learning and memory function. A circular pool with a diameter of 150 cm and a water temperature of 23±1°C was used, with non-toxic white pigment added to make the water surface opaque. A transparent escape platform (10 cm in diameter) was placed in the target quadrant, with the top of the platform 1.5 cm below the water surface.

[0206] Orientation navigation experiment (days 24-27): Tests were conducted from day 24 to day 28 after intervention (5 days in total). Four trials were conducted daily (at four different entry points). Rats entered the water facing the wall and swam freely to find the hidden platform, with a maximum allowed time of 90 seconds. If the platform was not found within 90 seconds, the experimenter guided the rat to the platform and allowed it to stay there for 15 seconds. The escape latency (in seconds) was recorded. A learning curve was plotted for a total of 16 trials over 4 days.

[0207] Space exploration experiment (day 28): On day 28, the platform was removed, and the rats entered the water from the fixed entry point and swam freely for 90 seconds. The ANY-maze video tracking system (or equivalent) recorded: percentage of time spent in the target quadrant (%), number of times the original platform position was crossed, and total swimming path length.

[0208] 4.4 Adhesive Removal Test

[0209] Assess sensorimotor function. Circular adhesive patches (10 mm in diameter) were applied to the palmar surfaces of both forepaws. Rats were placed in a transparent observation box, and the time of first contact with the patch on the affected side (contact latency, seconds) and the time of complete removal of the patch on the affected side (removal latency, seconds) were recorded. Each time point was tested three times, and the median value was used. The maximum permissible time was 120 seconds. Test time points: postoperative days 3, 7, 14, and 28.

[0210] 5. Assessment of cerebral infarction volume

[0211] Five additional rats from each group were sacrificed on day 28 after intervention for TTC staining (rats sampled independently of epigenetic age testing).

[0212] 5.1 TTC staining procedure

[0213] (1) After the rat was over-anesthetized, the head was cut off and the brain was removed. The fresh brain tissue was immediately placed in a -20°C freezer for 15 min (to make the tissue slightly hard and easier to slice).

[0214] (2) Using a rat brain mold, coronal sections of brain tissue were prepared with a thickness of 2 mm, and a total of about 6 to 7 sections were obtained.

[0215] (3) Place the slides in a 2% TTC solution (2,3,5-triphenyltetrazol, prepared with 0.1 M PBS, pH 7.4) preheated to 37°C and incubate at 37°C in the dark for 20 min, turning the slides over every 5 min to ensure uniform staining.

[0216] (4) Removed by TTC solution and fixed with 4% paraformaldehyde for 24 h.

[0217] (5) Take a picture of the stained brain slices with the front facing up (digital camera, fixed distance and lighting conditions).

[0218] 5.2 Quantitative analysis of infarct volume

[0219] Quantitative analysis was performed using ImageJ software (NIH, v1.53 or later): (1) For each brain slice photograph, delineate the area of ​​the healthy hemisphere (Ac), the area of ​​the affected hemisphere (Ai), and the area of ​​the infarct area on the affected side (Ainf, white area).

[0220] (2) The Swanson indirect method was used to correct the effect of cerebral edema: Corrected infarct area = Ac - (Ai - Ainf), that is, the area of ​​the healthy hemisphere minus the area of ​​the non-infarct area on the affected side.

[0221] (3) Infarct volume = Σ(corrected infarct area of ​​each section × slice thickness 2 mm).

[0222] (4) Percentage of infarct volume (%) = Infarct volume / volume of healthy hemisphere × 100%.

[0223] 6. Histological and Immunohistochemical / Immunofluorescence Detection

[0224] 6.1 Brain tissue fixation and section preparation

[0225] (1) On the 28th and 90th days after intervention (long-term safety observation), 5 rats in each group were fixed by cardiac perfusion. First, the rats were rapidly perfused with pre-cooled 0.9% NaCl until the outflow was clear (about 100 mL), and then slowly perfused with pre-cooled 4% paraformaldehyde (PFA, prepared with 0.1 M PB, pH 7.4) (200 mL, flow rate about 15 mL / min).

[0226] (2) After taking the whole brain, fix it in 4% PFA at 4°C overnight, and then dehydrate it in a gradient of 15% and 30% sucrose / PBS solution (each step until the tissue settles to the bottom).

[0227] (3) After OCT embedding, the sections were frozen (using a Leica CM1950 cryostat), with a section thickness of 20 μm (for immunofluorescence) or 30 μm (for stereotactic counting). Serial sections were collected and stored in cryoprotectant at -20°C.

[0228] (4) Take another part of the tissue for paraffin embedding (4% PFA fixation → 70% / 80% / 90% / 95% / 100% ethanol gradient dehydration → xylene clearing → paraffin impregnation → embedding), the paraffin section thickness is 5μm, for H&E staining and routine immunohistochemistry.

[0229] 6.2 Hematoxylin-eosin (H&E) staining

[0230] Paraffin sections were dewaxed with xylene (2 × 10⁻⁶ min) → rehydrated with a gradient of ethanol (100% → 95% → 80% → 70% → distilled water, 3 min each) → stained with hematoxylin and eosin for 5 min → rinsed with tap water and bluing for 5 min → differentiated with 1% hydrochloric acid alcohol for 3 s → rinsed with tap water → stained with eosin for 2 min → dehydrated with a gradient of ethanol → cleared with xylene → mounted with neutral resin. The tissue morphology was observed under an optical microscope, focusing on: the extent of the infarct area, changes in cell density in the ischemic penumbra, and the presence of abnormal proliferative foci or tumor-like structures.

[0231] 6.3 Neurogenesis-related immunofluorescence double labeling

[0232] It is used to detect newly generated neurons in the ischemic penumbra and SVZ region after intervention.

[0233] BrdU labeling regimen: BrdU administration regimen: 5-bromodeoxyuridine (BrdU, Sigma-Aldrich, B5002), 50 mg / kg, ip, bid, for 14 consecutive days during the intervention period (postoperative days 2-15) to label all dividing cells.

[0234] a. Wash frozen sections with PBS for 3 × 5 min. Block with PBS containing 10% normal donkey serum + 0.3% Triton X-100 at room temperature for 1 h.

[0235] b. BrdU antigen retrieval: BrdU antigen retrieval: DNA was denatured and exposed by treatment with 2 mol / L HCl at 37°C for 30 min, neutralized with 0.1 M sodium borate buffer (pH 8.5) for 10 min, and washed 3 times with PBS.

[0236] c. Primary antibody incubation (overnight at 4°C in a humidified chamber), Table 8.

[0237] Table 8 Antibody Incubation

[0238] d. After washing with PBS for 3×5 min, add the corresponding fluorescent secondary antibody: Donkey anti-Rat Alexa Fluor 594+Donkey anti-Rabbit Alexa Fluor 488 (or the corresponding host 488 / 594 combination), dilute 1:500, and incubate at room temperature in the dark for 1 h.

[0239] e. After washing with PBS, mount the slides with Fluoromount-G containing DAPI (Southern Biotech, 0100-20).

[0240] Confocal Imaging and Counting: Z-stack images (step size 1 μm, total thickness covering the full thickness of the slice) were acquired using a laser scanning confocal microscope (Zeiss LSM 880 or equivalent) under 20× and 40× objectives. Regions of interest (ROIs, area 0.5 mm² / field of view) were established in the ischemic penumbra cortex and the subventricular zone (SVZ), respectively. BrdU were counted in 6 sections on consecutive slices spaced 240 μm apart for each animal. + / DCX + and BrdU + / NeuN + Double-positive cell count. Double counting was avoided using optical dividers. Results are expressed as double-positive cells per mm².

[0241] 6.4 Angiogenesis Detection

[0242] Frozen sections were stained with CD31 (PECAM-1) immunofluorescence. Primary antibody: Goat anti-CD31 (1:200, R&D Systems AF3628), incubated overnight at 4°C. Secondary antibody: Donkey anti-Goat Alexa Fluor 594 (1:500), incubated for 1 h at room temperature. DAPI staining was also performed.

[0243] Microvessel density quantification: Quantitative method: Six fields of view (20× objective lens) were selected for each animal in the ischemic penumbra region, and CD31 was counted using ImageJ's Vessel Analysis plugin or manually. + For luminal structures (closed ring structures with an inner diameter > 5 μm are considered as one microvessel), calculate the microvessel density (MVD, vessels / mm²).

[0244] 6.5 Safety-related histopathological examination (brain tumor screening)

[0245] This test is a key endpoint for verifying the upper limit of the safe treatment window, focusing on assessing whether abnormal proliferation or tumor-like lesions occur in the G7 high-intensity group.

[0246] (1) H&E morphological screening: Two neuropathologists, who were blinded by grouping, independently reviewed serial whole-brain slides (one slide every 100 μm, covering the entire brain) on H&E sections taken on days 28 and 90. Assessment criteria: a. Areas with abnormally high cell density (>2 times the normal area); b. Increased nucleocytoplasmic ratio, nuclear atypia, and pathological mitotic figures; c. Pseudo-patrolitic necrosis (a characteristic feature of glioma); d. Microvascular proliferation.

[0247] Any suspected abnormal proliferative lesions were marked and transferred for immunohistochemical confirmation.

[0248] (2) Immunohistochemical confirmation panel: The following immunohistochemical markers were detected in the H&E screening positive areas, as shown in Table 9.

[0249] Table 9 Immunohistochemical marker information

[0250] Immunohistochemical procedure: After dewaxing and rehydration of paraffin sections, citrate antigen retrieval is performed (10 mM, pH 6.0, microwave on high for 10 min) → 3% H2O2 blocking of endogenous peroxidase for 10 min → 10% normal goat serum blocking for 30 min → primary antibody incubation at 4°C overnight → HRP-labeled secondary antibody (MaxVision HRP-Polymer anti-Mouse / Rabbit Kit, MXBBiotechnologies) at 37°C for 30 min → DAB staining (staining time monitored under a microscope) → hematoxylin counterstaining → dehydration, clearing, and mounting.

[0251] 6.6 Detection of genomic instability (γ-H2AX)

[0252] Frozen sections were stained with γ-H2AX (phospho-Histone H2A.X Ser139) immunofluorescence to detect DNA double-strand breaks (DSBs). Primary antibody: Mouse anti-γ-H2AX (1:500, Millipore 05-636, clone JBW301), incubated overnight at 4°C. Secondary antibody: Donkey anti-Mouse Alexa Fluor 488 (1:500), incubated for 1 h at room temperature. DAPI staining was also performed.

[0253] Quantitative method: 200 DAPIs were counted per animal in the ischemic penumbra region. + Cell nuclei were counted as the percentage of cells containing ≥5 γ-H2AX fluorescent focal points (foci). In normal brain tissue, this percentage is typically <3%; >10% indicates significant genomic instability.

[0254] 7. Molecular biological detection

[0255] 7.1 Detection of age-related secretory phenotype (SASP) biomarkers

[0256] Luminex Multifactor Assay: The level of SASP core factor in ischemic brain tissue homogenate was quantified using the Luminex / Milliplex cytokine assay panel (MilliporeRECYTMAG-65K). The detection factors include: IL-1β, IL-6, TNF-α, MCP-1 (CCL2), IL-8 (CXCL1, rat equivalent), MMP-3, MMP-9, and PAI-1 (Serpine1).

[0257] Procedure: Take 20 mg of ischemic penumbra tissue and homogenize it with RIPA lysis buffer containing a protease / phosphatase inhibitor cocktail. Centrifuge at 12,000×g for 15 min and collect the supernatant. After quantifying total protein using the BCA method, adjust to a uniform protein concentration (2 mg / mL) and analyze according to the Luminex kit instructions (Bio-Rad Bio-Plex 200). Results are expressed as pg / mg total protein.

[0258] 7.2 Detection of cell senescence markers

[0259] SA-β-gal staining: Frozen sections were stained for senescence-associated β-galactosidase (SA-β-gal). The Senescence β-Galactosidase Staining Kit (CST#9860) was used: sections were fixed with 4% PFA for 15 min → washed with PBS → incubated with X-gal staining solution at 37°C in the dark for 14 h (specifically detecting lysosomal β-gal activity in senescent cells at pH 6.0). Nuclear Fast Red counterstaining was then performed. Blue granule-positive cells were observed under a light microscope.

[0260] Western blot: Western blot detection of classic aging marker protein: p16 in ischemic penumbra brain tissue. TND4a (Rabbit,1:1000,Abcam ab211542), p21 Waf ¹ / cTp ¹(Rabbit, 1:1000, Abcam ab109520), γ-H2AX (same as above). GAPDH (1:5000) or β-actin (1:3000) were used as internal controls. Standard Western blot procedure: SDS-PAGE (12% gel) → wet transfer to PVDF membrane (100 V, 90 min) → blocking with 5% skim milk for 1 h → primary antibody overnight at 4°C → HRP secondary antibody at room temperature for 1 h → ECL chemiluminescence imaging (Bio-Rad ChemiDoc). ImageJ grayscale quantitative analysis.

[0261] 7.3 Validation of OSKGL transgenic expression (G4 / G5 / G7 groups only)

[0262] (1) RT-qPCR: RNA was collected from ischemic penumbra tissue (TRIzol extraction, DNase I treatment to remove genomic DNA contamination), and after reverse transcription, the mRNA levels of Oct4, Sox2, Klf4, Glis1, and Lin28 were detected by SYBR Green assay. Primers were designed to cross exon-exon junctions to exclude genomic DNA amplification. GAPDH was used as an internal control. - Quantification was performed using the ΔΔct method. Detection time points: during Dox administration (day 10) and after Dox withdrawal (days 21 and 28) to verify the controllability of the Tet-On system (high expression during Dox and silencing after withdrawal).

[0263] (2) Immunofluorescence: Oct4 antibody (Rabbit, 1:200, Abcam ab19857) + GFAP co-labeled to verify the expression of OSKGL protein in astrocytes.

[0264] 7.4 AAV Genome Integration Detection

[0265] LAM-PCR (Linear Amplification-Mediated PCR) was used to detect unexpected host chromosome integration into the AAV vector genome. This method can amplify host sequences flanking the insertion site without bias. Genomic DNA was extracted from the ischemic penumbra and linearly amplified using AAV ITR-specific primers, followed by adapter ligation, nested PCR, and sequencing. If integration events were detected, the integration frequency (number of integration events per million cells) and the integration site (whether it was located near a proto-oncogene or tumor suppressor gene) were recorded.

[0266] 8. Quantitative analysis of angiogenesis and neurotrophic factors

[0267] ELISA assay: The supernatant of homogenate from the ischemic penumbra tissue was used to quantify the following key factors (Table 10) using a commercially available ELISA kit: Table 10 Detection Factors

[0268] Procedure: Follow the standard double-antibody sandwich ELISA procedure according to the instructions for each kit. Samples should be prepared in duplicate wells. Read OD values ​​at 450 nm using a microplate reader (corrected for a 630 nm reference wavelength). Fit a standard curve using a four-parameter logistic regression and calculate the sample concentration. Results are expressed as pg / mg total protein.

[0269] 9. Statistical methods

[0270] All data are presented as mean ± standard deviation (Mean ± SD). Statistical analysis was performed using GraphPad Prism 10. (1) The independent samples t test is used for comparison between the two groups (when the data are normally distributed and have homogeneous variance).

[0271] (2) One-way ANOVA was used for comparisons among multiple groups, and Tukey HSD method (homogeneity of variance) or Games-Howell method (unequal variance) was used for post-hoc multiple comparisons.

[0272] (3) Repeated measures data (such as different time points of mNSS, water maze learning curve) were analyzed using two-way repeated measures ANOVA. Bonferroni post-hoc test was performed after the group × time interaction effect was significant.

[0273] (4) Fisher's exact test was used for count data such as tumor incidence.

[0274] (5) All tests were two-tailed, with a significance level set at α = 0.05. P < 0.05 was considered statistically significant, and P < 0.01 and P < 0.001 were labeled as statistically significant, respectively. and .

[0275] (6) Sample size estimation basis: Taking mNSS as the primary endpoint, assuming that the OSKGL group improved by 3 points in mNSS compared with the model group (SD=2.5), α=0.05, Power=0.80, the minimum sample size was calculated to be about 15 per group. Considering multi-endpoint detection, mid-term sampling and expected exclusion rate, 30 per group was set.

[0276] II. Results

[0277] 1. Quality control and preprocessing of RRBS sequencing raw data

[0278] D28 was the primary sampling time point, and GFAP was collected from the ischemic penumbra of 10 rats in each group. + Genomic DNA was extracted from astrocytes after FACS sorting, and RRBS library construction and NovaSeq PE150 sequencing were performed. Raw data from each sample were processed through the following standardized quality control pipeline.

[0279] 1.1 Sequencing data quality control standards are shown in Table 11.

[0280] Table 11 Sequencing Data Quality Control Standards

[0281] A complete quality control report for each sample is compiled and output by MultiQC. Unqualified samples are marked and excluded from clock analysis to ensure that the effective sample size for the final reversal amplitude calculation is ≥8 per group.

[0282] 1.2 Generation of the methylation β-value matrix

[0283] (1) Bismark (v0.24.0, -bowtie2 mode) aligned clean reads to the mouse reference genome mm10 / GRCm38. The methylation extractor module extracted the methylated and unmethylated read counts for each CpG site.

[0284] (2) Calculate the β value for each CpG: β = M / (M + U + α), where M = number of methylated reads, U = number of unmethylated reads, and α = 100 (pseudo-counting to prevent extreme fluctuations in β values ​​for low-coverage sites).

[0285] (3) Filtering: Remove CpG sites with a coverage depth <5×; remove SNP overlapping sites (dbSNP v151,MAF>0.01); remove sex chromosome CpG sites.

[0286] (4) Finally, a β value matrix is ​​generated (row = CpG site, column = sample). Each group has 800,000 to 1,200,000 effective CpGs, of which the average coverage of the 350 key CpGs required by the Horvath clock is 85% to 92%.

[0287] 1.3 Interpolation processing for missing CpG

[0288] For the few key clock-related CpG sites (8–15%) not covered by RRBS, k-nearest neighbor imputation (k=10) was used: based on the methylation patterns of other covered CpG sites in the sample, the 10 most similar samples were selected from the training dataset (Horvath's original mouse multi-tissue reference dataset), and the median β value of the corresponding missing CpG was used as the imputation value. This method has been validated in the literature (Petkovich et al., Cell Metab 2017), showing an impact of less than 1.5 weeks on the bias of clock-predicted age when the missing rate is <20%.

[0289] 2. Specific operations for calculating the Horvath epigenetic clock

[0290] 2.1 Clock Model Selection

[0291] The mouse-specific multi-tissue universal mouse DNA methylation clock model, published by Steve Horvath's laboratory, was used. This model uses 350 CpG loci selected by Elastic Net regression (α=0.5) as input features and outputs a predicted epigenetic age (in weeks) using a penalized regression formula. DNAmAge=Intercept+Σ i (β i ×w i ) Where β i w represents the methylation level of CpG at the i-th clock cycle. i The weights (positive or negative) represent the regression weights for the CpG, and the intercept represents the model intercept. All weights and intercept values ​​are public parameters.

[0292] 2.2 Computing Environment and Version Control

[0293] R version 4.2.0+. Load the mouse methylation clock R package (or equivalent implementation) released by Horvath Labs, using the latest stable release from the GitHub repository. All analysis code and package version numbers are recorded in the reproducibility log.

[0294] Input data format: β value matrix (CSV format, row name = CpG probe ID, column name = sample ID), supplemented by sample metadata file (containing information such as actual age, group, and gender).

[0295] 2.3 Sample-by-sample calculation process

[0296] (1) Load the β value matrix into the R environment and automatically match the intersection of the CpG required for the clock and the input data.

[0297] (2) Perform interpolation on the missing CpG.

[0298] (3) Call the clock prediction function to output the predicted epigenetic age (DNAmAge, weeks) for each sample.

[0299] (4) The actual age of the rats was 10-12 weeks (approximately 14 weeks at D28 after surgery), with the actual age of 14 weeks as the benchmark.

[0300] 3. Group-by-group calculation of aging acceleration value and reversal magnitude

[0301] 3.1 Definition and Calculation of Age Acceleration (AA)

[0302] AA = DNAmAge (predicted epigenetic age) - actual age in weeks; A positive value indicates that the cell population is "older than its actual age" at the epigenetic level, meaning accelerated aging has occurred. Astrocytes following ischemic injury are expected to exhibit a significantly positive AA value.

[0303] 3.2 The specific calculation of the AA value for each group is shown in Table 12.

[0304] Table 12 Results of age acceleration values ​​for each group

[0305] 3.3 Calculation of Reversal Magnitude (RM)

[0306] RM formula: Reversal magnitude (%) RM = [(Pre-intervention age acceleration value AA_baseline - Post-intervention age acceleration value AA_post) / Pre-intervention age acceleration value AA_baseline] × 100%; Where: AA_baseline = mean AA value of the G2 model group = +20.6 weeks. This represents the maximum epigenetic acceleration of aging caused by pMCAO, and serves as the unified denominator for calculating the reversal magnitude across all treatment groups.

[0307] AA_post = Measured AA value of each intervention group on D28.

[0308] The following is a step-by-step calculation process for the reversal magnitude of each group: G4:OSK Group AA_post = +13.2 weeks; RM=[(+20.6)-(+13.2)] / (+20.6)×100%=7.4 / 20.6×100%=35.9%; The reported percentage after rounding and SD propagation is 35% ± 8%.

[0309] Conclusion: RM=35%<57% (lower limit of the safety window), insufficient reversal.

[0310] G5:OSKGL Nucleoside

[0311] AA_post = +6.6 weeks; RM=[(+20.6)-(+6.6)] / |+20.6|×100%=14.0 / 20.6×100%=68.0%; The report states: 68% ± 7%; Conclusion: 57%≤RM=68%≤77%, accurately hitting the safe treatment window.

[0312] G6: OSKGL small molecule group

[0313] AA_post = +8.0 weeks; RM=[(+20.6)-(+8.0)] / |+20.6|×100%=12.6 / 20.6×100%=61.2%; The report states: 62% ± 9%; Conclusion: 57%≤RM=62%≤77%, accurately hitting the safe treatment window.

[0314] G7: High-strength OSKGL group

[0315] AA_post = +3.5 weeks; RM=[(+20.6)-(+3.5)] / |+20.6|×100%=17.1 / 20.6×100%=83.0%; The report states: 83% ± 11%; Conclusion: RM=83%>77% (safety window upper limit), excessive reversal, triggering a high-risk warning.

[0316] 3.4 SD Error Propagation Method

[0317] The reversal magnitude RM involves the ratio of two independent measurements (AA_baseline and AA_post), and the standard deviation is propagated using the Delta method (first-order Taylor expansion): SD_RM≈RM×√[(SD_AA_post / AA_diff)²+(SD_AA_baseline / AA_baseline)²] Where AA_diff = AA_baseline - AA_post. This approximation has good accuracy when the sample size is ≥10.

[0318] 4. Statistical methods for establishing the safe treatment window boundary (57%~77%)

[0319] The 57%–77% treatment window is not a prior hypothesis, but is established posteriorly from multidimensional experimental data using the following statistical methods.

[0320] 4.1 Window boundary exploration based on dose-response curve

[0321] In addition to the main experimental group, the preliminary dose exploration experiment included gradient AAV dose groups (0.5×, 1×, 2×, 3× standard dose) and gradient Dox induction time groups (7 days, 14 days, 21 days, 28 days), with n=5 in each group, for a total of 8 additional subgroups. GFAP was measured in each subgroup. + Horvath clock RM values ​​of astrocytes and their corresponding functional / safety endpoints.

[0322] Data from all subgroups and the main experimental group were integrated, and the following dose-response curves were plotted with RM as the independent variable: Curve 1: Improvement value of RM vs. mNSS (ΔmNSS = mNSS_G2 - mNSS_group) - efficacy curve; Curve 2: RM vs. Intracranial Abnormal Growth / Tumor Incidence (%) - Safety Curve; Curve 3: RM vs. BrdU + / NeuN + Double-positive cell density-neurogenesis curve.

[0323] 4.2 Establishment of the lower limit (57%)

[0324] In the efficacy curve (RM vs. ΔmNSS), when RM < 50%, there was no statistically significant difference in ΔmNSS between the G3 positive drug group and the G3 positive drug group (P > 0.05), indicating that the reprogramming intervention did not exceed the upper limit of efficacy of conventional treatment. When RM rose to approximately 57%, ΔmNSS showed a significant jump (inflection point), and thereafter, ΔmNSS further improved with increasing RM, but the slope gradually slowed down.

[0325] Specific statistical operations: Determination method: (1) Segmented / broken-stick regression was used to fit two linear models to the RM vs. ΔmNSS data to automatically identify the breakpoint. The significance of the breakpoint was verified by the Davies test using the segmented package (v1.6-4) in R.

[0326] (2) The estimated value of the inflection point and its 95% confidence interval are: 57.1% (95% CI: 52.3%~61.8%). The value is rounded down to 57% as the lower limit of the safety window.

[0327] (3) Biological validation: In all samples with RM < 57%, BrdU + / NeuN + The density of newborn neurons was not significantly different from that of the G2 model group, but the proportion of SA-β-gal positive cells was still significantly higher than that of the sham-operated group, indicating that senescent cells were not sufficiently cleared and endogenous neurogenesis was not effectively activated.

[0328] 4.3 Establishment of the upper limit (77%)

[0329] In the safety curve (RM vs. tumor incidence), when RM ≤ 77%, no abnormal proliferative foci were found in the brain H&E and Ki-67 tests of all samples, and the tumor incidence was 0%. When RM exceeded approximately 77%, sporadic abnormal proliferation began to appear, and the incidence increased sharply with further increases in RM.

[0330] Specific statistical operations: (1) Using tumor occurrence (presence / absence) as the binary dependent variable and RM as the continuous independent variable, fit a Logistic regression model: logit(P_tumor)=β0+β1×RM. Use the glm() function in R language.

[0331] (2) The RM threshold that makes the predicted tumor probability exceed 5% for the first time (i.e., the RM value when P_tumor=0.05) is determined as the statistical definition of the upper limit of the safety window. The estimated value of this threshold and its 95% CI are: 76.8% (95% CI: 72.1%~81.5%). Rounded to 77%.

[0332] (3) Auxiliary verification - Receiver Operating Characteristic (ROC) curve analysis: ROC curves were constructed to predict tumor occurrence using RM>threshold. The Youden index (sensitivity + specificity - 1) reached its maximum value when the threshold was 77%, with AUC=0.94 (95% CI: 0.87~1.00), indicating that 77% has excellent discrimination ability.

[0333] 4.4 Confirmation of the optimal therapeutic range within the window

[0334] Within the 57%–77% window, further analysis was conducted on the relationship between RM and various efficacy indicators: (1) The improvement of mNSS reached a plateau in the RM 60%~70% range (ΔmNSS≈5~6 points). After continuing to rise to 77%, ΔmNSS did not improve significantly further (P>0.05 vs. 70% group), but the inter-individual variation increased.

[0335] (2) BrdU + / NeuN + The density peaks in the range of RM 65%~75%.

[0336] (3) Comprehensive benefit-risk analysis: RM 60%~70% is the optimal range. G5 (68%) and G6 (62%) both fall within this range, which is consistent with theoretical expectations.

[0337] 5. Statistical validation of the data

[0338] 5.1 Statistical test of differences in RM between groups

[0339] One-way ANOVA was used to test whether there were significant differences in the RM values ​​of the four groups G4 / G5 / G6 / G7. After the ANOVA was significant (F(3,36)=45.7, P<0.001), Tukey HSD post-hoc multiple comparisons were performed. The comparison methods and results are shown in Table 13.

[0340] Table 13 Comparison methods and results

[0341] Key statistical conclusions: a) There was no significant difference in reversal rates between G5 and G6 (P=0.52), indicating that the two delivery pathways of nucleic acid and small molecules can achieve equivalent reversal rates; b) G4 (35%) was significantly lower than G5 / G6 (68% / 62%), P<0.001, directly proving the necessity of Glis1 and Lin28 for reaching the therapeutic window; c) G7 (83%) was significantly higher than G5 / G6, confirming that high dose + prolonged induction led to excessive reversal.

[0342] 5.2 Determining the position of each group of RM and the boundary of the safety window

[0343] A one-sample t-test was performed on each group of RM values ​​to test whether their means were significantly different from 57% (lower limit) and 77% (upper limit), respectively. The test results are shown in Table 14.

[0344] Table 14 Test Results

[0345] Note: For G6, the mean was 62%, and although the p-value (vs. 57%) was 0.14 (not <0.05), 8 out of 10 samples had individual RM values ​​>57%, and the group mean was in the middle of the window. For G7, the mean was 83% (vs. 77%), with a p-value of 0.12, but 7 out of 10 samples had values ​​>77%, and the functional / safety data consistently supported the overreversal decision. Therefore, the in-window / out-of-window decision comprehensively considers point estimation, individual distribution, and functional endpoint data.

[0346] 6. Multi-cell type cross-validation

[0347] To verify GFAP + The clock results for astrocytes are not cell type-specific false positives; NeuN cells synchronously sorted from the same brain tissue... + Neurons perform parallel Horvath clock analysis.

[0348] 6.1 NeuN + The magnitude of neuronal epigenetic age reversal

[0349] The results are shown in Table 15.

[0350] Table 15. Magnitude of Reversal in Neuronal Epigenetic Age

[0351] The Pearson correlation coefficient between the two cell types for RM was r=0.96 (P<0.001, n=40, all intervention groups combined), indicating a high degree of consistency in Horvath clock detection results across different neuronal cell types. The systematic RM of astrocytes was slightly higher than that of neurons (approximately 4–7%), possibly related to the more active proliferation / remodeling characteristics of astrocytes after stroke, but the trend was entirely consistent.

[0352] This cross-validation demonstrates that: a) the concept of a safe treatment window is universally applicable across major brain cell types; and b) using astrocytes as a representative assay is both biologically sound and technically feasible.

[0353] 7. Clock dynamics verification at the intermediate time point (D14)

[0354] To assess the temporal dynamics of epigenetic age reversal, samples were collected in advance at mid-D14 timepoint (n=5 in each group) for Horvath clock detection. This data validated the process monitoring function of the clock. The results are shown in Table 16.

[0355] Table 16 Verification Results

[0356] Closed-loop control implications: Key finding: At D14, the RM of G7 has reached 70% (close to the 77% upper limit). If clock detection is initiated at this point and an early warning is issued, and the Dox is removed in time, theoretically, the final RM of G7 can be controlled within the window. This data directly supports the technical feasibility of the closed-loop control system—real-time clock monitoring can provide a sufficient intervention window before excessive reversal actually occurs.

[0357] 8. Comprehensive judgment criteria

[0358] The DNA methylation clock determination for each group was not based solely on a mechanical comparison of RM values, but rather on a systematic determination that integrated information from the following four dimensions, as shown in Table 17. The comprehensive assessment results of the magnitude of epigenetic age reversal in astrocytes after intervention are shown in Table 18.

[0359] Table 17 Judgment Criteria

[0360] Table 18 Comprehensive Judgment Results

[0361] Only when all four criteria are met simultaneously—Dimension 1 (RM within the window), Dimension 2 (most individuals fall within the window), Dimension 3 (significant functional improvement), and Dimension 4 (zero tumor risk)—is the treatment window considered precisely targeted, safe, and effective. G5 and G6 meet all four criteria simultaneously; G4 does not meet Dimensions 1 and 3 (insufficient reversal and no functional improvement); G7 does not meet Dimensions 1 and 4 (excessive reversal and tumor development).

[0362] 9. Clock detection repeatability verification

[0363] 9.1 Technical repeatability (repeated testing of the same sample)

[0364] Six samples were randomly selected (one sample per group), and the DNA from each sample was divided in half. RRBS library construction and sequencing were performed independently, and the DNA mAge was calculated for each half. The intra-group correlation coefficient (ICC) between technical replicates was 0.97 (95% CI: 0.92–0.99), and the mean difference in DNA mAge was 0.8 ± 0.6 weeks, indicating that the detection system has excellent technical repeatability.

[0365] 9.2 Biological repeatability (consistency among different individuals in the same group)

[0366] Within-group coefficients of variation (CV): G1=11.8%, G2=10.1%, G4=15.3%, G5=10.3%, G6=14.5%, G7=13.3%. All CVs were <20%, indicating acceptable biological reproducibility. The slightly higher CVs for G4 and G6 may reflect the effects of the incomplete three-factor protocol and individual variability in small molecule pharmacokinetics, respectively.

[0367] 9.3 Consistency of positive / negative internal controls

[0368] (1) Negative control: The DNAmAge of the G1 sham-operated group was highly consistent with the actual age (difference < 2 weeks), which verified the accuracy of the clock model calibration in this experimental system.

[0369] (2) Positive control: The G2 model group showed consistent and significant accelerated aging (AA=+20.6 weeks, P<0.001 vs. G1), which verified that the pMCAO model did indeed induce epigenetic accelerated aging that could be detected by the clock.

[0370] (3) The successful internal controls verified the sensitivity (detection of ischemia-induced aging) and specificity (no false positives in the non-damage group) of the Horvath clock under the experimental conditions, providing a reliable anchor for the interpretation of the treatment group data.

[0371] 10. Logical Association Verification

[0372] 10.1 Correlation Analysis between Clock Prediction and mNSS Improvement

[0373] Linear regression analysis was performed using the RM value of each animal individual as the independent variable and D28 mNSS as the dependent variable. Results: β = -0.12 (for every 1% increase in RM, mNSS decreased by 0.12 points), R² = 0.68, P < 0.001. This indicates that RM can explain 68% of the mNSS variation, and the two are strongly negatively correlated—the greater the magnitude of reversal, the better the recovery of neurological function (within the window).

[0374] 10.2 Discriminant Analysis of Clock Prediction and Tumor Occurrence

[0375] Using RM>77% as the positive criterion for clock warning and D28 / D90 histopathological confirmation of tumor / dysplasia as the gold standard, the diagnostic efficacy of the clock warning system was calculated. The indicators are shown in Table 19 and the results are shown in Table 20.

[0376] Table 19 Detection Indicators

[0377] Table 20 Validation Results

[0378] The NPV of 100% is the most important statistical support for the safety claim of this invention: in all 54 samples with a RM ≤ 77%, not a single case showed any intracranial tumor or abnormal growth at day 28 or day 90. This means that as long as the treatment regimen keeps the RM within a safe window, the risk of brain tumors can be reduced to zero.

[0379] A PPV of 45% (not 100%) means that a RM > 77% does not necessarily lead to tumors—some individuals with excessive reversal may not develop overt tumors due to protective mechanisms such as host immune surveillance. However, from a safety perspective, this invention adopts a conservative strategy: any signal of RM > 77% is considered a high-risk warning and triggers intervention to reduce or stop the treatment.

[0380] The consistency between Table 18 and Table 20 forms the experimental basis for the predictive function of the detection system of the present invention.

[0381] Table 21 D28 neurological deficit scores (mNSS) for each group

[0382] The scoring range was 0–18 points, with 0 points representing normal and 18 points representing the most severe deficit. The D1 score served as the baseline for each group, and there was no statistically significant difference between groups (one-way ANOVA, p=0.92). Statistical methods included repeated measures two-way ANOVA and post-hoc Tukey HSD test. The D28 standard deviation in group G7 was significantly larger than in other groups (Levene test, p<0.01), reflecting significant inter-individual variability in treatment efficacy.

[0383] Conclusions: The D28 mNSS scores of G5 (OSKGL nucleic acid group) and G6 (OSKGL small molecule group) improved by 5.4 and 4.6 points respectively compared with the G2 model group (both p<0.001), which were significantly better than those of the G3 edaravone positive control group (improvement of 2.6 points, p<0.05). The improvement in the G4 OSK group was only 1.2 points and not statistically significant, confirming the necessity of the five-factor OSKGL. The average improvement of the G7 high-intensity group was similar to that of G5, but the individual differences were huge (SD=3.2), suggesting that excessive reversal leads to unstable efficacy. The above mNSS results were strongly negatively correlated with the magnitude of Horvath clock reversal (β=-0.12, R²=0.68, p<0.001), verifying that the degree of epigenetic age reversal can predict the degree of neurological function recovery.

[0384] Table 22. Cerebral infarction volume on day 28 (quantitative TTC staining) in each group

[0385] TTC staining was performed at the end of day 28, with coronal sections of brain tissue (2 mm thick, 6 sections / mouse). White areas represented infarcted tissue, and red areas represented viable tissue. Infarct volume was corrected for edema using the Swanson indirect method: Corrected infarct volume = contralateral hemisphere volume - (ipsilateral hemisphere volume - infarct volume). Statistical analysis was performed using one-way ANOVA plus Tukey HSD post-hoc test.

[0386] Conclusion: The infarct volume in the G5 OSKGL nucleic acid group was reduced by 52% compared to the G2 model group, and by 45% in the G6 small molecule group, both significantly better than the G3 edaravone group (reduced by 23%). The G7 high-intensity group had the smallest infarct volume (reduced by 61%), but also the largest standard deviation (28.4 mm³), consistent with the mNSS results, reflecting increased inter-individual variability due to excessive reversal.

[0387] Table 23 Results of intracranial safety histopathological examination of each group on day 28. p<0.01 vs G2 (Fisher exact test); γ-H2AX: p<0.001 vs G2 (Tukey HSD) All 210 rats underwent systematic whole-brain serial section H&E staining at either the end of observation (D28) or the extended observation point (D90) (5 additional rats from the G7 group were included for the extended observation period to D90). Abnormal proliferative foci were defined as areas of abnormally high cell density (>2 times normal) visible under H&E staining, confirmed by Ki67 immunohistochemistry with a proliferation index >20%. Glioma-like lesions were defined as GFAP-positive, high-density lesions exhibiting nuclear atypia and angiogenesis. The γ-H2AX positive cell rate was expressed as the percentage of γ-H2AX nuclear focal positive cells per high-power field (400×) in the ischemic penumbra, with six fields counted per animal and the average value taken.

[0388] Conclusion: No abnormal proliferation or tumor-like lesions were detected in the brains of any of the 120 rats in G1–G6 (incidence rate 0%). Abnormal proliferation was only observed in 5 cases (25%) of the G7 high-intensity group, including 2 cases of glioma-like lesions. This result is completely consistent with the Horvath clock warning: in the G7 group, RM = 83% > 77% (upper limit of the safety window), the clock issued a high-risk signal on day 28, and histopathological examination on day 90 confirmed tumor formation. The RMs of the G5 and G6 groups were 68% and 62%, respectively, both falling within the 57%–77% safety window, corresponding to a zero incidence of brain tumors. γ-H2AX detection results showed a significantly increased genomic instability in the G7 group, further supporting the safety risks caused by excessive reprogramming.

[0389] Table 24 Results of aging marker detection in the ischemic penumbra of each group on day 28. p<0.001 vs G2 (Tukey HSD).

[0390] The relative expression levels of p16^INK4a and p21^Cip1 mRNA were normalized to 1.0 with the G1 sham-operated group as the baseline. RT-qPCR was used for detection, with β-actin as the internal control. The SA-β-gal positive cell rate was expressed as the percentage of blue-positive cells per high-power field (200×) of X-gal staining in frozen sections of the ischemic penumbra. IL-6 and TNF-α were quantified using Luminex multifactor assay (Bio-Rad Bio-Plex 200), in pg / mg total protein.

[0391] Conclusion: pMCAO injury led to a comprehensive increase in senescence markers in the ischemic penumbra (p16 increased 8.5-fold in the G2 group, and the SA-β-gal positivity rate reached 42.5%), confirming the pathological basis of "accelerated senescence after ischemia". Both the G5 OSKGL nucleic acid group and the G6 small molecule group reversed senescence markers to near-sham-operated levels (p16 decreased by 2.2-fold and 2.8-fold, respectively, and SA-β-gal decreased by 12.4% and 15.6%, respectively), and SASP factors (IL-6 and TNF-α) also decreased significantly, confirming that both regimens achieved neuroprotection by reversing cellular senescence. The G7 high-intensity group showed the largest decrease in senescence markers, but considering its 25% brain tumor incidence, this indicates that excessive reversal, while maximally eliminating senescent cells, also exceeded safety limits. The G3 edaravone group and the G4 OSK group only partially reduced senescence markers, consistent with their limited Horvath clock reversal.

[0392] Table 25 ELISA quantification of angiogenesis and neurotrophic factors in the ischemic penumbra of each group on day 28 (pg / mg total protein) p<0.001 vs G2 (Tukey HSD) Conclusions: In the G2 model group, VEGF-A increased compensatorily after pMCAO injury, but the vascular maturation factor Ang-1 and neurotrophic factors (BDNF, NGF, GDNF) decreased significantly, reflecting insufficient angiogenesis and lack of neurotrophic support after ischemia. The G5OSKGL nucleic acid group and the G6 small molecule group not only further upregulated VEGF-A (promoting angiogenesis) but also restored Ang-1 to levels above the sham-operated level (promoting the maturation and stabilization of new blood vessels). Simultaneously, BDNF, NGF, and GDNF all significantly increased, suggesting that partial reprogramming of OSKGL restarted the endogenous neurovascular repair program in the ischemic area by reversing cellular senescence. Neurotrophic factors in the G3 edaravone group and the G4 OSK group only showed a slight increase, consistent with their limited epigenetic reversal.

[0393] Table 26. Quantitative analysis of microvessel density and neurogenesis in the ischemic penumbra of each group on day 28. p<0.001 vs G2 (Tukey HSD) CD31 + Microvessel density was quantified using the ImageJ Vessel Analysis plugin in the ischemic penumbra region at 20× field of view (closed CD31 with an inner diameter >5μm). + (Lumen structure), average of 6 fields of view for each animal. BrdU + / DCX + (newborn neural progenitor cells) and BrdU + / NeuN + Double-positive cells (newborn and mature neurons) were counted in the ischemic penumbra cortex and subventricular zone (SVZ) using a laser confocal microscope (40× objective, Z-stack). The average value was taken from 6 sections per animal. BrdU was injected intraperitoneally for labeling (50 mg / kg, bid×7 days) 7 days before sampling.

[0394] Conclusion: After pMCAO injury, the microvessel density in the G2 model group decreased by 47% compared to the sham-operated group, reflecting vascular loss in the ischemic area. The microvessel density in the G5 OSKGL nucleic acid group recovered to levels above the sham-operated level (212.6 vs 185.4 vessels / mm²), suggesting that OSKGL intervention not only protects surviving vessels but also promotes angiogenesis, consistent with the upregulation of VEGF-A and Ang-1 in Table 25. Regarding neurogenesis, the BrdU... + / DCX + Cellular compensatory increase but BrdU + / NeuN + The absence of an increase in mature neurons indicates that although neural progenitors proliferate after injury, their differentiation and maturation are impaired. BrdU in groups G5 and G6 + / NeuN + The number of cells was 5.1 times and 4.2 times that of the G2 group, respectively, confirming that OSKGL partial reprogramming significantly promoted the differentiation and maturation of newborn neurons, providing a cellular basis for functional recovery.

[0395] As demonstrated by the above embodiments, the application of the detection and treatment window theory based on the DNA methylation clock (Horvath clock) in this invention shows that only the reversal amplitude of groups G5 and G6 precisely falls within the safety window of 57% to 77%. The final neurological function and safety endpoint data fully validated the accuracy of this detection; the group within the window achieved optimal functional recovery and was absolutely safe; the group outside the window was either ineffective or at high risk. This proves that this invention not only provides an effective therapeutic composition but also a set of "molecular metrics" and "risk control systems" that have been validated by authoritative research to ensure its safe and precise application.

[0396] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A brain-targeting composition, characterized in that, Including brain-targeting nucleic acid compositions or brain-penetrating small molecule compositions; The brain-targeting nucleic acid composition includes the Oct4 gene, Sox2 gene, Klf4 gene, Glis1 gene, and Lin28 gene; The brain-penetrating small molecule composition comprises the following components at the following concentrations: 150-250 μM valproic acid, 6-10 μM CHIR99021, 6-10 μM Repsox-E616452, 3-5 μM transphenylcyclopropane, and 30-50 μM forscolin.

2. The brain-targeting composition according to claim 1, characterized in that, The genes in the brain-targeting nucleic acid composition are obtained by tandem 2A peptide sequences.

3. Use of the brain-targeting composition according to claim 1 or 2 in the preparation of a medicament for the treatment and / or prevention of stroke.

4. A brain-targeting expression cassette, characterized in that, Includes any one of the following: (1) Expression vector: The brain-targeting nucleic acid composition is linked into a delivery vector; (2) Recombinant cells: Brain-targeting nucleic acid composition was transfected into HEK293T cells; The brain-targeting nucleic acid composition is the brain-targeting nucleic acid composition as described in claim 1 or 2.

5. The brain-targeting expression cassette according to claim 4, characterized in that, The delivery vector is a recombinant adeno-associated virus or modified lipid nanoparticles; The recombinant adeno-associated virus is AAV-PHP.eB or AAV9.

6. The use of the brain-targeting expression cassette according to claim 4 or 5 in the preparation of medicaments for the treatment and / or prevention of stroke.

7. A method for evaluating the effect of a non-disease-diagnostic drug on stroke repair, characterized in that, Includes the following steps: (I) Select brain tissue before and after drug intervention and sort out the target cells; (II) Extract genomic DNA from the target cells, perform DNA methylation profile analysis, and obtain methylation data; (III) Input the methylation data into the Horvath epigenetic clock model to obtain the predicted epigenetic age of the target cells; (IV) Substitute the predicted epigenetic age into the formula to calculate the reversal magnitude; (V) Evaluate the repair effect of drugs on stroke based on the degree of reversal.

8. The evaluation method according to claim 7, characterized in that, The drug is the brain-targeting composition according to claim 1 or 2; The target cells include one or more of neurons, astrocytes, oligodendrocyte precursor cells, and endogenous neural stem cells.

9. The evaluation method according to claim 7, characterized in that, The DNA methylation profiling analysis method is either whole-genome bisulfite sequencing or high-precision methylation chip method.

10. The evaluation method according to claim 7, characterized in that, The formula for calculating the reversal magnitude is as follows: Reversal magnitude % = [(Pre-intervention age acceleration value - Post-intervention age acceleration value) / Pre-intervention age acceleration value] × 100%; Age acceleration value = predicted epigenetic age - actual age; The best effect on stroke repair is achieved when the reversal rate is 57% to 77%.

Images